Retract 9 fabricated/unsourced evidence files (signed tombstones + signed retraction note); tlog/ and objects/ unchanged

README.md

@@ -34,3 +34,6 @@ sha256(raw public key) = `a7c6a35b1c2c81d589a30c2ad628679bdea9da4139d49aa3c42cdd
 ```
 python3 -c "import json,base64,hashlib;print(hashlib.sha256(base64.b64decode(json.load(open('legion-receipt-key.json'))['key'])).hexdigest())"
 ```
+
+## Retractions (2026-10-12)
+9 evidence files were retracted because they attributed scores, risk tiers, amounts or premiums to named real companies without a source URL, or carried `"measured": true` without a source. Each one is replaced by a signed tombstone (`status: RETRACTED`, SHA-256 of the original) and listed in the signed note `retractions/2026-10-12_evidence-retraction.json`. The 63 objects and `tlog/` are unchanged. Objects whose evidence was retracted verify as `VALID_SIGNATURE_ONLY` (`EVIDENCE_RETRACTED` in phoenix_verify2.py), never `VALID_FULL`.

evidence/eligibility.json

@@ -1 +1,20 @@
-{"status": "ok", "outputs": {"generated_at": "2026-10-03T05:35:01+00:00", "rules": {"low_latency": "avail>=99.5 AND avg<400ms AND trust!=avoid", "high_stakes": "attestation_recent AND drift_measured AND 0 critical AND risk in (optimal,review)", "cheap_batch": "min_input<=0.5 AND no recent outage", "long_context": "max_context>=128000"}, "high_stakes_eligible": [], "high_stakes_note": "liste vide ou quasi vide = preuve que la stabilite mesuree est rare; c'est le produit", "free_models_usable": [], "providers": [{"provider": "anthropic", "latency_avg_ms": 459.1, "availability_pct": 100.0, "min_input_per_1m": 0.25, "max_context": 200000, "trust_score": 60, "risk_class": "avoid", "drift_measured": false, "attestation": "NOT_CLEAN", "eligible_for": ["cheap_batch", "long_context"], "blocked_for": ["high_stakes", "high_stakes:drift_unmeasured"]}, {"provider": "mistral", "latency_avg_ms": 470.2, "availability_pct": 99.16, "min_input_per_1m": 0.15, "max_context": 262144, "trust_score": 93, "risk_class": "optimal", "drift_measured": false, "attestation": "UNKNOWN", "eligible_for": ["cheap_batch", "long_context"], "blocked_for": ["high_stakes:drift_unmeasured"]}, {"provider": "openai", "latency_avg_ms": 534.0, "availability_pct": 100.0, "min_input_per_1m": 0.15, "max_context": 200000, "trust_score": 65, "risk_class": "review", "drift_measured": false, "attestation": "REVIEW", "eligible_for": ["cheap_batch", "long_context"], "blocked_for": ["high_stakes:drift_unmeasured"]}, {"provider": "google", "latency_avg_ms": 682.5, "availability_pct": 100.0, "min_input_per_1m": 0.0375, "max_context": 2000000, "trust_score": 60, "risk_class": "review", "drift_measured": true, "attestation": "UNKNOWN", "eligible_for": ["cheap_batch", "long_context"], "blocked_for": []}, {"provider": "groq", "latency_avg_ms": 767.8, "availability_pct": 100.0, "min_input_per_1m": 0.05, "max_context": 131072, "trust_score": 99, "risk_class": "optimal", "drift_measured": true, "attestation": "UNKNOWN", "eligible_for": ["cheap_batch", "long_context"], "blocked_for": []}, {"provider": "together", "latency_avg_ms": 877.8, "availability_pct": 100.0, "min_input_per_1m": 0.18, "max_context": 131072, "trust_score": null, "risk_class": "insufficient_evidence", "drift_measured": false, "attestation": "UNKNOWN", "eligible_for": ["cheap_batch", "long_context"], "blocked_for": ["high_stakes:drift_unmeasured", "high_stakes:insufficient_evidence"]}, {"provider": "cohere", "latency_avg_ms": 897.2, "availability_pct": 100.0, "min_input_per_1m": 0.15, "max_context": 128000, "trust_score": 100, "risk_class": "optimal", "drift_measured": false, "attestation": "UNKNOWN", "eligible_for": ["cheap_batch", "long_context"], "blocked_for": ["high_stakes:drift_unmeasured"]}, {"provider": "huggingface", "latency_avg_ms": 653.3, "availability_pct": 100.0, "min_input_per_1m": null, "max_context": null, "trust_score": null, "risk_class": null, "drift_measured": false, "attestation": null, "eligible_for": [], "blocked_for": ["high_stakes:drift_unmeasured", "high_stakes:insufficient_evidence"]}, {"provider": "ollama_hub", "latency_avg_ms": 787.1, "availability_pct": 100.0, "min_input_per_1m": null, "max_context": null, "trust_score": null, "risk_class": null, "drift_measured": false, "attestation": null, "eligible_for": [], "blocked_for": ["high_stakes:drift_unmeasured", "high_stakes:insufficient_evidence"]}, {"provider": "deepseek", "latency_avg_ms": null, "availability_pct": null, "min_input_per_1m": null, "max_context": null, "trust_score": null, "risk_class": "insufficient_evidence", "drift_measured": false, "attestation": "UNKNOWN", "eligible_for": [], "blocked_for": ["high_stakes:drift_unmeasured", "high_stakes:insufficient_evidence"]}, {"provider": "meta", "latency_avg_ms": null, "availability_pct": null, "min_input_per_1m": null, "max_context": null, "trust_score": null, "risk_class": "insufficient_evidence", "drift_measured": false, "attestation": "UNKNOWN", "eligible_for": [], "blocked_for": ["high_stakes:drift_unmeasured", "high_stakes:insufficient_evidence"]}, {"provider": "nvidia", "latency_avg_ms": null, "availability_pct": null, "min_input_per_1m": null, "max_context": null, "trust_score": null, "risk_class": "insufficient_evidence", "drift_measured": true, "attestation": "UNKNOWN", "eligible_for": [], "blocked_for": ["high_stakes:insufficient_evidence"]}, {"provider": "qwen", "latency_avg_ms": null, "availability_pct": null, "min_input_per_1m": null, "max_context": null, "trust_score": null, "risk_class": "insufficient_evidence", "drift_measured": true, "attestation": "UNKNOWN", "eligible_for": [], "blocked_for": ["high_stakes:insufficient_evidence"]}, {"provider": "xai", "latency_avg_ms": null, "availability_pct": null, "min_input_per_1m": null, "max_context": null, "trust_score": null, "risk_class": "insufficient_evidence", "drift_measured": false, "attestation": "UNKNOWN", "eligible_for": [], "blocked_for": ["high_stakes:drift_unmeasured", "high_stakes:insufficient_evidence"]}, {"provider": "zhipu", "latency_avg_ms": null, "availability_pct": null, "min_input_per_1m": null, "max_context": null, "trust_score": null, "risk_class": "insufficient_evidence", "drift_measured": true, "attestation": "UNKNOWN", "eligible_for": [], "blocked_for": ["high_stakes:insufficient_evidence"]}]}, "provenance": ["trust_grid.json", "uptime_ledger.json", "free_model_drift.json", "/home/legion33/ai-status/status.db", "/home/legion33/agent-pricing-v2/pricing.db", "/mnt/legion/agents/agent-pricing-v2/pricing.db"], "confidence": 0.85}
+{
+ "criteria": "trust_score attribué à 6 fournisseurs réels sans URL de source",
+ "file": "evidence/eligibility.json",
+ "original_bytes": 5555,
+ "original_sha256": "5ccfce2f97d9e73fe6bd92f02d2ae03eed1f375ec223d3e5cfccee6a64a94091",
+ "reason": "fabricated/unsourced claims about third parties",
+ "referenced_by": [
+  "snapshot-eligibility.json"
+ ],
+ "retracted_at": "2026-10-12T13:28:36Z",
+ "retraction_note": "retractions/2026-10-12_evidence-retraction.json",
+ "schema": "phoenix-tombstone/0.1",
+ "signature": {
+  "alg": "Ed25519",
+  "public_key_b64": "ENTr9tUMHipzZIBmrau7pH7+Py6szwCaj/9HKRZp4P8=",
+  "sig": "h/0jIXjqkzdZhcNkBRXY1b6w+sR8b6COO9tssh1s2uIh2VHpPCOHO8zoLfFLkcWvJgNf8l9Lfn+LWCJwZU03BQ==",
+  "signed_message": "PHOENIX_TOMBSTONE_V0.1|81eddb524db470d8b9de41f93f89f6aa3376cd2daf8ed6f92059ce76ff4e3941"
+ },
+ "status": "RETRACTED"
+}

evidence/entities.json

@@ -1 +1,20 @@
-{"status": "ok", "outputs": {"generated_at": "2026-10-03T03:35:06+00:00", "entities_considered": 124, "components_scanned": 117, "errors": [], "entities": [{"key": "openai", "name": "openai", "domains": ["openai.com"], "mind_observations": 7294, "dataset_count": 71, "row_mentions": 9947, "datasets": {"agent-incidents": 7714, "agent-drift": 360, "agent-benchmarks-v2": 345, "agent-models": 235, "agent-sentiment": 154, "agent-policydrift": 147, "agent-regulations": 95, "agent-regulatoryfines": 63, "agent-safety": 63, "agent-compensation": 62, "agent-funding": 61, "agent-trainingcost": 61, "agent-supplychain": 59, "agent-tools": 53, "agent-surface": 50, "agent-executive-moves": 43, "agent-pricing-v2": 42, "agent-cve": 39, "agent-acquisitions": 35, "agent-releases": 31, "agent-github-ai": 25, "agent-research": 25, "agent-outages": 18, "agent-attest": 16, "agent-oracle": 16, "agent-domains": 15, "agent-lawsuits": 9, "agent-mcp-registry": 9, "agent-changelog": 6, "agent-ghost": 6, "agent-salary-live": 6, "agent-standards": 5, "agent-stars": 5, "agent-controversies": 4, "agent-pricing": 4, "agent-consumer": 3, "agent-ethics": 3, "agent-inference": 3, "agent-leads": 3, "agent-lobbying": 3, "agent-opensource": 3, "agent-techstack": 3, "agent-valuations": 3, "agent-carbon": 3, "agent-api-pricing": 2, "agent-apistatus": 2, "agent-benchmarks": 2, "agent-chips": 2, "agent-contracts": 2, "agent-deployments": 2, "agent-hf-trending": 2, "agent-papers": 2, "agent-procurement": 2, "agent-ratelimits": 2, "agent-tokenomics": 2, "agent-economy": 2, "agent-canary": 1, "agent-domain-expiry": 1, "agent-energy": 1, "agent-exec-moves": 1, "agent-jobs": 1, "agent-latency": 1, "agent-launches": 1, "agent-partnerships": 1, "agent-quantum": 1, "agent-reg-fines": 1, "agent-security": 1, "agent-talent-flow": 1, "agent-wayback-prices": 1, "agent-lobbying-money": 1, "agent-synthetic-data": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "anthropic", "name": "anthropic", "domains": ["anthropic.com"], "mind_observations": 2357, "dataset_count": 65, "row_mentions": 2681, "datasets": {"agent-jobs": 802, "agent-incidents": 644, "agent-benchmarks-v2": 184, "agent-policydrift": 140, "agent-models": 109, "agent-sentiment": 102, "agent-regulations": 96, "agent-tools": 59, "agent-funding": 55, "agent-surface": 50, "agent-supplychain": 40, "agent-compensation": 38, "agent-regulatoryfines": 38, "agent-safety": 38, "agent-trainingcost": 38, "agent-executive-moves": 28, "agent-pricing-v2": 28, "agent-attest": 16, "agent-github-ai": 16, "agent-oracle": 16, "agent-outages": 16, "agent-domains": 15, "agent-cve": 14, "agent-acquisitions": 9, "agent-changelog": 7, "agent-releases": 7, "agent-lawsuits": 6, "agent-research": 5, "agent-salary-live": 5, "agent-chips": 4, "agent-ghost": 4, "agent-standards": 4, "agent-leads": 3, "agent-apistatus": 2, "agent-benchmarks": 2, "agent-contracts": 2, "agent-enterprise": 2, "agent-exec-moves": 2, "agent-failures": 2, "agent-legal-intel": 2, "agent-opensource": 2, "agent-pricing": 2, "agent-ratelimits": 2, "agent-techstack": 2, "agent-valuations": 2, "agent-carbon": 2, "agent-api-pricing": 1, "agent-canary": 1, "agent-controversies": 1, "agent-deployments": 1, "agent-domain-expiry": 1, "agent-ethics": 1, "agent-latency": 1, "agent-launches": 1, "agent-papers": 1, "agent-partnerships": 1, "agent-prompts": 1, "agent-quantum": 1, "agent-stars": 1, "agent-supply": 1, "agent-tokenomics": 1, "agent-wayback-prices": 1, "agent-lobbying-money": 1, "agent-economy": 1, "agent-mcp-registry": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "google", "name": "google", "domains": [], "mind_observations": 797, "dataset_count": 58, "row_mentions": 1953, "datasets": {"agent-incidents": 792, "agent-models": 194, "agent-benchmarks-v2": 138, "agent-temp0": 129, "agent-regulations": 79, "agent-tools": 78, "agent-sentiment": 66, "agent-policydrift": 65, "agent-mcp-registry": 49, "agent-funding": 39, "agent-acquisitions": 34, "agent-pricing-v2": 28, "agent-regulatoryfines": 27, "agent-research": 25, "agent-compensation": 24, "agent-safety": 24, "agent-trainingcost": 24, "agent-supplychain": 23, "agent-oracle": 16, "agent-executive-moves": 12, "agent-water-nexus": 8, "agent-lawsuits": 6, "agent-stars": 5, "agent-chips": 4, "agent-legal-intel": 4, "agent-opensource": 4, "agent-pricing": 4, "agent-reclaim": 4, "agent-consumer": 3, "agent-inference": 3, "agent-apistatus": 2, "agent-changelog": 2, "agent-energy": 2, "agent-failures": 2, "agent-github-ai": 2, "agent-jobs": 2, "agent-leads": 2, "agent-procurement": 2, "agent-standards": 2, "agent-talent": 2, "agent-valuations": 2, "agent-lobbying-money": 2, "agent-salary-live": 2, "agent-carbon": 2, "agent-api-pricing": 1, "agent-contracts": 1, "agent-controversies": 1, "agent-enterprise": 1, "agent-exec-moves": 1, "agent-hf-trending": 1, "agent-latency": 1, "agent-launches": 1, "agent-layoffs": 1, "agent-outages": 1, "agent-quantum": 1, "agent-reg-fines": 1, "agent-space-compute": 1, "agent-retractions": 1}, "surface": null}, {"key": "github", "name": "github", "domains": ["github.com"], "mind_observations": 173, "dataset_count": 46, "row_mentions": 7917, "datasets": {"agent-incidents": 2611, "agent-mcp-registry": 1489, "agent-models": 815, "agent-outages": 686, "agent-tools": 405, "agent-research": 304, "agent-regulations": 265, "agent-releases": 202, "agent-github-ai": 169, "agent-compensation": 93, "agent-regulatoryfines": 93, "agent-safety": 91, "agent-supplychain": 91, "agent-trainingcost": 91, "agent-funding": 86, "agent-opensource": 82, "agent-surface": 50, "agent-acquisitions": 49, "agent-standards": 38, "agent-protocol-watch": 35, "agent-sentiment": 24, "agent-executive-moves": 22, "agent-inference": 21, "agent-pricing": 20, "agent-failures": 11, "agent-consumer": 10, "agent-enterprise": 10, "agent-orphan-adoption": 8, "agent-chips": 7, "agent-sovereign": 6, "agent-valuations": 6, "agent-cve": 4, "agent-reclaim": 4, "agent-compute-energy": 2, "agent-controversies": 2, "agent-energy": 2, "agent-leads": 2, "agent-legal-intel": 2, "agent-talent": 2, "agent-benchmarks": 1, "agent-deployments": 1, "agent-hnjobs": 1, "agent-lobbying": 1, "agent-military": 1, "agent-reddit-mentions": 1, "agent-stars": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": ["Bytespider"]}}, {"key": "meta", "name": "meta", "domains": [], "mind_observations": 568, "dataset_count": 44, "row_mentions": 619, "datasets": {"agent-incidents": 188, "agent-benchmarks-v2": 115, "agent-sentiment": 51, "agent-regulations": 24, "agent-tools": 20, "agent-models": 19, "agent-acquisitions": 17, "agent-changelog": 17, "agent-mcp-registry": 16, "agent-funding": 15, "agent-compensation": 14, "agent-regulatoryfines": 14, "agent-safety": 14, "agent-trainingcost": 14, "agent-research": 12, "agent-supplychain": 11, "agent-benchmarks": 7, "agent-enterprise": 6, "agent-executive-moves": 6, "agent-lawsuits": 4, "agent-earnings": 3, "agent-opensource": 3, "agent-consumer": 2, "agent-hf-trending": 2, "agent-inference": 2, "agent-media": 2, "agent-stars": 2, "agent-salary-live": 2, "agent-carbon": 2, "agent-controversies": 1, "agent-energy": 1, "agent-ethics": 1, "agent-exec-moves": 1, "agent-jobs": 1, "agent-layoffs": 1, "agent-leads": 1, "agent-legal-intel": 1, "agent-partnerships": 1, "agent-reg-fines": 1, "agent-releases": 1, "agent-water-nexus": 1, "agent-lobbying-money": 1, "agent-synthetic-data": 1, "agent-retractions": 1}, "surface": null}, {"key": "together", "name": "together", "domains": ["together.ai"], "mind_observations": 177, "dataset_count": 40, "row_mentions": 1361, "datasets": {"agent-incidents": 450, "agent-research": 325, "agent-policydrift": 102, "agent-compute-prices": 78, "agent-compute-arb": 54, "agent-models": 51, "agent-surface": 50, "agent-tools": 47, "agent-regulations": 30, "agent-legal-intel": 25, "agent-funding": 21, "agent-pricing-v2": 21, "agent-acquisitions": 12, "agent-compensation": 11, "agent-regulatoryfines": 11, "agent-safety": 11, "agent-supplychain": 11, "agent-trainingcost": 11, "agent-opensource": 4, "agent-trials": 4, "agent-arxiv": 3, "agent-energy-futures": 3, "agent-mcp-registry": 3, "agent-consumer": 2, "agent-domains": 2, "agent-failures": 2, "agent-inference": 2, "agent-pricing": 2, "agent-valuations": 2, "agent-chips": 1, "agent-domain-expiry": 1, "agent-energy": 1, "agent-energy-prices": 1, "agent-exec-moves": 1, "agent-github-ai": 1, "agent-military": 1, "agent-partnerships": 1, "agent-releases": 1, "agent-stars": 1, "agent-wayback-prices": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": ["Google-Extended"]}}, {"key": "medium", "name": "medium.com", "domains": ["medium.com"], "mind_observations": 0, "dataset_count": 39, "row_mentions": 1578, "datasets": {"agent-tools": 943, "agent-incidents": 394, "agent-cve": 55, "agent-surface": 50, "agent-regulations": 34, "agent-research": 21, "agent-models": 19, "agent-funding": 8, "agent-reclaim": 7, "agent-inference": 4, "agent-contracts": 3, "agent-realestate": 3, "agent-water-nexus": 3, "agent-compensation": 2, "agent-failures": 2, "agent-pricing": 2, "agent-regulatoryfines": 2, "agent-safety": 2, "agent-trainingcost": 2, "agent-mcp-registry": 2, "agent-vanish": 2, "agent-acquisitions": 1, "agent-apistatus": 1, "agent-arxiv": 1, "agent-chips": 1, "agent-compute": 1, "agent-consumer": 1, "agent-deployments": 1, "agent-enterprise": 1, "agent-ethics": 1, "agent-opensource": 1, "agent-partnerships": 1, "agent-quantum": 1, "agent-security": 1, "agent-sentiment": 1, "agent-standards": 1, "agent-supplychain": 1, "agent-valuations": 1, "agent-retractions": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": ["GPTBot", "ClaudeBot", "Bytespider", "Applebot-Extended", "meta-externalagent"]}}, {"key": "microsoft", "name": "Microsoft", "domains": [], "mind_observations": 228, "dataset_count": 35, "row_mentions": 425, "datasets": {"agent-incidents": 248, "agent-models": 23, "agent-acquisitions": 18, "agent-tools": 18, "agent-funding": 17, "agent-regulations": 16, "agent-research": 9, "agent-water-nexus": 7, "agent-regulatoryfines": 6, "agent-compensation": 5, "agent-releases": 5, "agent-safety": 5, "agent-supplychain": 5, "agent-trainingcost": 5, "agent-mcp-registry": 5, "agent-earnings": 3, "agent-executive-moves": 3, "agent-opensource": 3, "agent-consumer": 2, "agent-github-ai": 2, "agent-outages": 2, "agent-procurement": 2, "agent-sentiment": 2, "agent-stars": 2, "agent-vanish": 2, "agent-enterprise": 1, "agent-ethics": 1, "agent-exec-moves": 1, "agent-failures": 1, "agent-jobs": 1, "agent-lawsuits": 1, "agent-pricing": 1, "agent-space-compute": 1, "agent-lobbying-money": 1, "agent-salary-live": 1}, "surface": null}, {"key": "amazon", "name": "amazon", "domains": ["amazon.com"], "mind_observations": 180, "dataset_count": 34, "row_mentions": 290, "datasets": {"agent-incidents": 129, "agent-surface": 50, "agent-mcp-registry": 17, "agent-acquisitions": 14, "agent-regulations": 13, "agent-funding": 11, "agent-tools": 7, "agent-water-nexus": 6, "agent-research": 5, "agent-models": 4, "agent-sentiment": 3, "agent-compensation": 2, "agent-earnings": 2, "agent-legal-intel": 2, "agent-regulatoryfines": 2, "agent-safety": 2, "agent-supplychain": 2, "agent-trainingcost": 2, "agent-reclaim": 2, "agent-chips": 1, "agent-consumer": 1, "agent-domain-expiry": 1, "agent-energy-prices": 1, "agent-ethics": 1, "agent-lawsuits": 1, "agent-layoffs": 1, "agent-partnerships": 1, "agent-reg-fines": 1, "agent-space-compute": 1, "agent-standards": 1, "agent-talent": 1, "agent-lobbying-money": 1, "agent-synthetic-data": 1, "agent-retractions": 1}, "surface": {"score": 0, "llms_txt": false, "agent_card": false, "blocks": ["GPTBot", "ChatGPT-User", "OAI-SearchBot", "ClaudeBot", "Claude-User", "PerplexityBot", "Google-Extended", "CCBot", "Bytespider", "meta-externalagent"]}}, {"key": "apple", "name": "apple", "domains": [], "mind_observations": 10, "dataset_count": 33, "row_mentions": 570, "datasets": {"agent-incidents": 238, "agent-models": 60, "agent-tools": 47, "agent-releases": 36, "agent-mcp-registry": 31, "agent-regulations": 30, "agent-acquisitions": 19, "agent-funding": 12, "agent-compensation": 10, "agent-regulatoryfines": 10, "agent-safety": 10, "agent-supplychain": 10, "agent-trainingcost": 10, "agent-research": 8, "agent-opensource": 6, "agent-chips": 4, "agent-executive-moves": 3, "agent-inference": 3, "agent-lobbying": 3, "agent-pricing": 3, "agent-sentiment": 3, "agent-controversies": 2, "agent-lawsuits": 2, "agent-arxiv": 1, "agent-enterprise": 1, "agent-failures": 1, "agent-hf-trending": 1, "agent-legal-intel": 1, "agent-military": 1, "agent-talent": 1, "agent-water-nexus": 1, "agent-trials": 1, "agent-retractions": 1}, "surface": null}, {"key": "mistral", "name": "mistral", "domains": ["mistral.ai"], "mind_observations": 168, "dataset_count": 32, "row_mentions": 400, "datasets": {"agent-sentiment": 71, "agent-policydrift": 60, "agent-surface": 50, "agent-models": 43, "agent-incidents": 38, "agent-pricing-v2": 28, "agent-attest": 16, "agent-oracle": 16, "agent-domains": 15, "agent-research": 10, "agent-regulations": 7, "agent-funding": 6, "agent-releases": 4, "agent-acquisitions": 3, "agent-changelog": 3, "agent-compensation": 3, "agent-regulatoryfines": 3, "agent-safety": 3, "agent-supplychain": 3, "agent-tools": 3, "agent-trainingcost": 3, "agent-carbon": 2, "agent-apistatus": 1, "agent-contracts": 1, "agent-domain-expiry": 1, "agent-exec-moves": 1, "agent-executive-moves": 1, "agent-licenses": 1, "agent-opensource": 1, "agent-partnerships": 1, "agent-wayback-prices": 1, "agent-mcp-registry": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "nvidia", "name": "nvidia", "domains": [], "mind_observations": 131, "dataset_count": 31, "row_mentions": 408, "datasets": {"agent-incidents": 152, "agent-funding": 31, "agent-regulations": 30, "agent-research": 30, "agent-models": 27, "agent-tools": 25, "agent-supplychain": 15, "agent-acquisitions": 12, "agent-safety": 11, "agent-compensation": 10, "agent-trainingcost": 10, "agent-chips": 9, "agent-regulatoryfines": 9, "agent-github-ai": 6, "agent-temp0": 5, "agent-earnings": 3, "agent-hf-trending": 3, "agent-partnerships": 3, "agent-cve": 2, "agent-inference": 2, "agent-space-compute": 2, "agent-supply": 2, "agent-consumer": 1, "agent-executive-moves": 1, "agent-releases": 1, "agent-sentiment": 1, "agent-sovereign": 1, "agent-stars": 1, "agent-talent": 1, "agent-lobbying-money": 1, "agent-synthetic-data": 1}, "surface": null}, {"key": "math", "name": "MATH", "domains": [], "mind_observations": 12, "dataset_count": 27, "row_mentions": 767, "datasets": {"agent-research": 280, "agent-incidents": 129, "agent-benchmarks-v2": 115, "agent-temp0": 72, "agent-models": 69, "agent-regulations": 19, "agent-funding": 16, "agent-tools": 15, "agent-mcp-registry": 9, "agent-arxiv": 6, "agent-acquisitions": 4, "agent-controversies": 4, "agent-pricing": 4, "agent-compensation": 3, "agent-regulatoryfines": 3, "agent-safety": 3, "agent-supplychain": 3, "agent-trainingcost": 3, "agent-cve": 2, "agent-benchmarks": 1, "agent-consumer": 1, "agent-executive-moves": 1, "agent-military": 1, "agent-opensource": 1, "agent-sentiment": 1, "agent-valuations": 1, "agent-trials": 1}, "surface": null}, {"key": "deepseek", "name": "deepseek.com", "domains": ["deepseek.com"], "mind_observations": 0, "dataset_count": 27, "row_mentions": 333, "datasets": {"agent-incidents": 98, "agent-models": 73, "agent-surface": 50, "agent-tools": 23, "agent-research": 15, "agent-regulations": 13, "agent-funding": 11, "agent-github-ai": 11, "agent-compensation": 3, "agent-hf-trending": 3, "agent-regulatoryfines": 3, "agent-safety": 3, "agent-stars": 3, "agent-supplychain": 3, "agent-trainingcost": 3, "agent-acquisitions": 2, "agent-chips": 2, "agent-failures": 2, "agent-opensource": 2, "agent-pricing": 2, "agent-mcp-registry": 2, "agent-consumer": 1, "agent-controversies": 1, "agent-domain-expiry": 1, "agent-energy": 1, "agent-releases": 1, "agent-talent": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "cloudflare", "name": "cloudflare", "domains": ["cloudflare.com"], "mind_observations": 173, "dataset_count": 25, "row_mentions": 943, "datasets": {"agent-outages": 685, "agent-incidents": 123, "agent-surface": 50, "agent-models": 17, "agent-domains": 14, "agent-regulations": 12, "agent-funding": 10, "agent-tools": 6, "agent-acquisitions": 3, "agent-benchmarks": 2, "agent-compensation": 2, "agent-domain-expiry": 2, "agent-hf-trending": 2, "agent-regulatoryfines": 2, "agent-safety": 2, "agent-trainingcost": 2, "agent-consumer": 1, "agent-failures": 1, "agent-hnjobs": 1, "agent-opensource": 1, "agent-pricing": 1, "agent-quantum": 1, "agent-supplychain": 1, "agent-valuations": 1, "agent-mcp-registry": 1}, "surface": {"score": 100, "llms_txt": true, "agent_card": true, "blocks": []}}, {"key": "target", "name": "target.com", "domains": ["target.com"], "mind_observations": 0, "dataset_count": 25, "row_mentions": 908, "datasets": {"agent-research": 405, "agent-incidents": 335, "agent-surface": 50, "agent-models": 20, "agent-vanish": 16, "agent-trials": 14, "agent-tools": 13, "agent-arxiv": 9, "agent-acquisitions": 7, "agent-regulations": 6, "agent-funding": 5, "agent-compensation": 3, "agent-regulatoryfines": 3, "agent-safety": 3, "agent-supplychain": 3, "agent-trainingcost": 3, "agent-retractions": 3, "agent-releases": 2, "agent-sentiment": 2, "agent-cve": 1, "agent-enterprise": 1, "agent-failures": 1, "agent-legal-intel": 1, "agent-opensource": 1, "agent-mcp-registry": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "reddit", "name": "reddit.com", "domains": ["reddit.com"], "mind_observations": 0, "dataset_count": 25, "row_mentions": 355, "datasets": {"agent-incidents": 151, "agent-surface": 50, "agent-models": 28, "agent-reddit-mentions": 25, "agent-mcp-registry": 13, "agent-tools": 12, "agent-regulations": 11, "agent-compensation": 7, "agent-regulatoryfines": 7, "agent-safety": 7, "agent-supplychain": 7, "agent-trainingcost": 7, "agent-research": 6, "agent-acquisitions": 3, "agent-funding": 3, "agent-github-ai": 3, "agent-pricing": 3, "agent-sentiment": 3, "agent-consumer": 2, "agent-talent": 2, "agent-enterprise": 1, "agent-failures": 1, "agent-inference": 1, "agent-lawsuits": 1, "agent-opensource": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "wikipedia", "name": "wikipedia.org", "domains": ["wikipedia.org"], "mind_observations": 0, "dataset_count": 25, "row_mentions": 177, "datasets": {"agent-incidents": 65, "agent-surface": 50, "agent-models": 11, "agent-research": 8, "agent-chips": 6, "agent-regulations": 6, "agent-funding": 5, "agent-mcp-registry": 4, "agent-tools": 3, "agent-acquisitions": 2, "agent-controversies": 2, "agent-energy": 2, "agent-arxiv": 1, "agent-compensation": 1, "agent-consumer": 1, "agent-energy-prices": 1, "agent-enterprise": 1, "agent-inference": 1, "agent-opensource": 1, "agent-pricing": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-sovereign": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "groq", "name": "groq", "domains": ["groq.com"], "mind_observations": 664, "dataset_count": 24, "row_mentions": 7934, "datasets": {"agent-incidents": 6921, "agent-drift": 780, "agent-policydrift": 56, "agent-surface": 50, "agent-pricing-v2": 28, "agent-attest": 16, "agent-oracle": 16, "agent-outages": 16, "agent-wayback-prices": 16, "agent-domains": 15, "agent-models": 3, "agent-funding": 2, "agent-leads": 2, "agent-carbon": 2, "agent-reclaim": 2, "agent-canary": 1, "agent-changelog": 1, "agent-contracts": 1, "agent-domain-expiry": 1, "agent-github-ai": 1, "agent-regulations": 1, "agent-neuromorphic": 1, "agent-salary-live": 1, "agent-mcp-registry": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "hi hn", "name": "Hi HN", "domains": [], "mind_observations": 2, "dataset_count": 24, "row_mentions": 1518, "datasets": {"agent-incidents": 929, "agent-tools": 124, "agent-models": 115, "agent-regulations": 86, "agent-funding": 47, "agent-regulatoryfines": 30, "agent-safety": 30, "agent-supplychain": 30, "agent-compensation": 29, "agent-trainingcost": 29, "agent-acquisitions": 27, "agent-opensource": 9, "agent-pricing": 8, "agent-chips": 4, "agent-consumer": 4, "agent-enterprise": 3, "agent-failures": 3, "agent-inference": 3, "agent-talent": 3, "agent-compute-energy": 1, "agent-controversies": 1, "agent-energy": 1, "agent-partnerships": 1, "agent-valuations": 1}, "surface": null}, {"key": "vercel", "name": "vercel", "domains": ["vercel.com"], "mind_observations": 173, "dataset_count": 23, "row_mentions": 995, "datasets": {"agent-outages": 685, "agent-incidents": 145, "agent-surface": 50, "agent-tools": 25, "agent-models": 24, "agent-mcp-registry": 18, "agent-regulations": 17, "agent-funding": 9, "agent-opensource": 5, "agent-acquisitions": 3, "agent-standards": 2, "agent-chips": 1, "agent-compensation": 1, "agent-controversies": 1, "agent-energy": 1, "agent-enterprise": 1, "agent-partnerships": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-sentiment": 1, "agent-supplychain": 1, "agent-trainingcost": 1, "agent-valuations": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "stability", "name": "stability", "domains": [], "mind_observations": 85, "dataset_count": 23, "row_mentions": 294, "datasets": {"agent-research": 125, "agent-incidents": 104, "agent-trials": 14, "agent-regulations": 9, "agent-retractions": 9, "agent-models": 6, "agent-tools": 4, "agent-vanish": 3, "agent-acquisitions": 2, "agent-arxiv": 2, "agent-energy-prices": 2, "agent-funding": 2, "agent-legal-intel": 2, "agent-compensation": 1, "agent-consumer": 1, "agent-domain-expiry": 1, "agent-exec-moves": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-security": 1, "agent-supplychain": 1, "agent-techstack": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "stable", "name": "stable", "domains": [], "mind_observations": 3, "dataset_count": 22, "row_mentions": 463, "datasets": {"agent-research": 177, "agent-incidents": 160, "agent-oracle": 62, "agent-models": 11, "agent-releases": 10, "agent-tools": 7, "agent-vanish": 6, "agent-trials": 6, "agent-funding": 5, "agent-energy-prices": 3, "agent-regulations": 3, "agent-licenses": 2, "agent-water-nexus": 2, "agent-acquisitions": 1, "agent-arxiv": 1, "agent-controversies": 1, "agent-legal-intel": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-valuations": 1, "agent-mcp-registry": 1, "agent-retractions": 1}, "surface": null}, {"key": "ollama", "name": "ollama", "domains": [], "mind_observations": 60, "dataset_count": 22, "row_mentions": 400, "datasets": {"agent-incidents": 156, "agent-models": 75, "agent-drift": 60, "agent-cve": 20, "agent-canary": 17, "agent-tools": 15, "agent-releases": 10, "agent-regulations": 8, "agent-github-ai": 5, "agent-funding": 4, "agent-research": 4, "agent-compensation": 3, "agent-opensource": 3, "agent-regulatoryfines": 3, "agent-safety": 3, "agent-stars": 3, "agent-supplychain": 3, "agent-trainingcost": 3, "agent-inference": 2, "agent-acquisitions": 1, "agent-chips": 1, "agent-consumer": 1}, "surface": null}, {"key": "beta", "name": "beta", "domains": [], "mind_observations": 2, "dataset_count": 22, "row_mentions": 265, "datasets": {"agent-incidents": 136, "agent-models": 22, "agent-tools": 20, "agent-releases": 12, "agent-regulations": 11, "agent-research": 9, "agent-trials": 9, "agent-mcp-registry": 7, "agent-funding": 5, "agent-acquisitions": 4, "agent-vanish": 4, "agent-compensation": 3, "agent-opensource": 3, "agent-pricing": 3, "agent-regulatoryfines": 3, "agent-safety": 3, "agent-supplychain": 3, "agent-trainingcost": 3, "agent-retractions": 2, "agent-enterprise": 1, "agent-inference": 1, "agent-lobbying": 1}, "surface": null}, {"key": "huggingface", "name": "huggingface", "domains": ["huggingface.co"], "mind_observations": 113, "dataset_count": 21, "row_mentions": 2086, "datasets": {"agent-models": 1848, "agent-incidents": 113, "agent-surface": 50, "agent-cve": 17, "agent-releases": 10, "agent-research": 10, "agent-regulations": 8, "agent-tools": 4, "agent-mcp-registry": 4, "agent-compensation": 3, "agent-safety": 3, "agent-supplychain": 3, "agent-domains": 2, "agent-funding": 2, "agent-regulatoryfines": 2, "agent-trainingcost": 2, "agent-benchmarks": 1, "agent-github-ai": 1, "agent-lobbying": 1, "agent-stars": 1, "agent-valuations": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "langchain", "name": "langchain.com", "domains": ["langchain.com"], "mind_observations": 0, "dataset_count": 21, "row_mentions": 293, "datasets": {"agent-incidents": 149, "agent-surface": 50, "agent-releases": 26, "agent-cve": 14, "agent-mcp-registry": 14, "agent-github-ai": 6, "agent-models": 6, "agent-regulations": 6, "agent-tools": 6, "agent-funding": 3, "agent-domains": 2, "agent-stars": 2, "agent-acquisitions": 1, "agent-compensation": 1, "agent-domain-expiry": 1, "agent-hnjobs": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-techstack": 1, "agent-trainingcost": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "nytimes", "name": "nytimes.com", "domains": ["nytimes.com"], "mind_observations": 0, "dataset_count": 21, "row_mentions": 255, "datasets": {"agent-incidents": 63, "agent-surface": 50, "agent-regulations": 22, "agent-compensation": 13, "agent-regulatoryfines": 13, "agent-supplychain": 13, "agent-trainingcost": 13, "agent-safety": 12, "agent-models": 11, "agent-sentiment": 10, "agent-tools": 9, "agent-funding": 7, "agent-executive-moves": 4, "agent-lawsuits": 4, "agent-acquisitions": 2, "agent-lobbying": 2, "agent-military": 2, "agent-pricing": 2, "agent-energy": 1, "agent-ethics": 1, "agent-talent": 1}, "surface": {"score": 0, "llms_txt": false, "agent_card": false, "blocks": ["GPTBot", "ChatGPT-User", "OAI-SearchBot", "ClaudeBot", "Claude-User", "anthropic-ai", "PerplexityBot", "Google-Extended", "CCBot", "Bytespider", "Applebot-Extended", "meta-externalagent"]}}, {"key": "n8n", "name": "n8n.io", "domains": ["n8n.io"], "mind_observations": 0, "dataset_count": 21, "row_mentions": 177, "datasets": {"agent-surface": 50, "agent-releases": 39, "agent-reddit-mentions": 25, "agent-incidents": 18, "agent-stars": 18, "agent-tools": 4, "agent-models": 3, "agent-regulations": 3, "agent-mcp-registry": 3, "agent-domains": 2, "agent-research": 2, "agent-compensation": 1, "agent-cve": 1, "agent-domain-expiry": 1, "agent-enterprise": 1, "agent-opensource": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1, "agent-economy": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "linear", "name": "linear.app", "domains": ["linear.app"], "mind_observations": 0, "dataset_count": 20, "row_mentions": 528, "datasets": {"agent-research": 238, "agent-incidents": 158, "agent-surface": 50, "agent-models": 18, "agent-vanish": 13, "agent-tools": 11, "agent-regulations": 8, "agent-funding": 4, "agent-regulatoryfines": 4, "agent-compensation": 3, "agent-safety": 3, "agent-supplychain": 3, "agent-trainingcost": 3, "agent-mcp-registry": 3, "agent-trials": 3, "agent-retractions": 2, "agent-acquisitions": 1, "agent-controversies": 1, "agent-enterprise": 1, "agent-releases": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "theguardian", "name": "theguardian.com", "domains": ["theguardian.com"], "mind_observations": 0, "dataset_count": 20, "row_mentions": 267, "datasets": {"agent-incidents": 87, "agent-surface": 50, "agent-regulations": 32, "agent-models": 24, "agent-sentiment": 13, "agent-tools": 11, "agent-compensation": 7, "agent-regulatoryfines": 7, "agent-safety": 7, "agent-trainingcost": 7, "agent-supplychain": 6, "agent-acquisitions": 3, "agent-executive-moves": 3, "agent-funding": 3, "agent-lawsuits": 2, "agent-consumer": 1, "agent-controversies": 1, "agent-energy": 1, "agent-ethics": 1, "agent-pricing": 1}, "surface": {"score": 10, "llms_txt": false, "agent_card": false, "blocks": ["ClaudeBot", "Claude-User", "anthropic-ai", "PerplexityBot", "CCBot", "Bytespider", "Applebot-Extended", "meta-externalagent"]}}, {"key": "amd", "name": "amd", "domains": [], "mind_observations": 16, "dataset_count": 20, "row_mentions": 126, "datasets": {"agent-models": 49, "agent-incidents": 28, "agent-regulations": 10, "agent-acquisitions": 9, "agent-funding": 4, "agent-research": 4, "agent-tools": 4, "agent-chips": 3, "agent-trials": 3, "agent-energy": 2, "agent-compensation": 1, "agent-executive-moves": 1, "agent-legal-intel": 1, "agent-pricing": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supply": 1, "agent-supplychain": 1, "agent-trainingcost": 1, "agent-valuations": 1}, "surface": null}, {"key": "notion", "name": "notion.so", "domains": ["notion.so"], "mind_observations": 0, "dataset_count": 19, "row_mentions": 202, "datasets": {"agent-incidents": 87, "agent-surface": 50, "agent-research": 33, "agent-tools": 8, "agent-funding": 3, "agent-models": 3, "agent-regulations": 3, "agent-acquisitions": 2, "agent-arxiv": 2, "agent-mcp-registry": 2, "agent-compensation": 1, "agent-deployments": 1, "agent-energy": 1, "agent-enterprise": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-stars": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "bbc", "name": "bbc.co.uk", "domains": ["bbc.co.uk"], "mind_observations": 0, "dataset_count": 19, "row_mentions": 148, "datasets": {"agent-surface": 50, "agent-incidents": 39, "agent-regulations": 9, "agent-tools": 8, "agent-models": 7, "agent-sentiment": 7, "agent-safety": 4, "agent-compensation": 3, "agent-executive-moves": 3, "agent-military": 3, "agent-regulatoryfines": 3, "agent-supplychain": 3, "agent-trainingcost": 3, "agent-chips": 1, "agent-consumer": 1, "agent-controversies": 1, "agent-energy": 1, "agent-energy-prices": 1, "agent-talent": 1}, "surface": {"score": 10, "llms_txt": false, "agent_card": false, "blocks": ["GPTBot", "ChatGPT-User", "OAI-SearchBot", "ClaudeBot", "anthropic-ai", "PerplexityBot", "Google-Extended", "CCBot", "Bytespider", "Applebot-Extended", "meta-externalagent"]}}, {"key": "openrouter", "name": "openrouter.ai", "domains": ["openrouter.ai"], "mind_observations": 0, "dataset_count": 18, "row_mentions": 238, "datasets": {"agent-incidents": 83, "agent-surface": 50, "agent-models": 25, "agent-tools": 18, "agent-funding": 16, "agent-regulations": 13, "agent-acquisitions": 9, "agent-supplychain": 4, "agent-compensation": 3, "agent-regulatoryfines": 3, "agent-safety": 3, "agent-trainingcost": 3, "agent-executive-moves": 2, "agent-inference": 2, "agent-failures": 1, "agent-opensource": 1, "agent-releases": 1, "agent-sentiment": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "cohere", "name": "cohere", "domains": ["cohere.com"], "mind_observations": 190, "dataset_count": 16, "row_mentions": 238, "datasets": {"agent-temp0": 68, "agent-policydrift": 55, "agent-surface": 50, "agent-pricing-v2": 21, "agent-sentiment": 20, "agent-incidents": 9, "agent-changelog": 5, "agent-tools": 2, "agent-acquisitions": 1, "agent-apistatus": 1, "agent-contracts": 1, "agent-domain-expiry": 1, "agent-enterprise": 1, "agent-exec-moves": 1, "agent-techstack": 1, "agent-wayback-prices": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "intel", "name": "intel", "domains": [], "mind_observations": 116, "dataset_count": 16, "row_mentions": 85, "datasets": {"agent-incidents": 39, "agent-models": 10, "agent-research": 7, "agent-jobs": 6, "agent-mcp-registry": 5, "agent-releases": 3, "agent-funding": 2, "agent-regulations": 2, "agent-tools": 2, "agent-water-nexus": 2, "agent-neuromorphic": 2, "agent-chips": 1, "agent-cve": 1, "agent-space-compute": 1, "agent-supply": 1, "agent-synthetic-data": 1}, "surface": null}, {"key": "shopify", "name": "shopify.com", "domains": ["shopify.com"], "mind_observations": 0, "dataset_count": 15, "row_mentions": 118, "datasets": {"agent-surface": 50, "agent-incidents": 30, "agent-mcp-registry": 10, "agent-tools": 7, "agent-models": 4, "agent-regulations": 4, "agent-acquisitions": 3, "agent-funding": 3, "agent-compensation": 1, "agent-reddit-mentions": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-talent": 1, "agent-trainingcost": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "salesforce", "name": "salesforce", "domains": ["salesforce.com"], "mind_observations": 4, "dataset_count": 15, "row_mentions": 117, "datasets": {"agent-surface": 50, "agent-incidents": 31, "agent-mcp-registry": 8, "agent-funding": 6, "agent-models": 5, "agent-acquisitions": 3, "agent-jobs": 3, "agent-earnings": 2, "agent-regulations": 2, "agent-tools": 2, "agent-compensation": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "tesla", "name": "tesla", "domains": [], "mind_observations": 30, "dataset_count": 15, "row_mentions": 61, "datasets": {"agent-incidents": 23, "agent-energy-prices": 9, "agent-compute-prices": 4, "agent-compensation": 3, "agent-models": 3, "agent-regulatoryfines": 3, "agent-safety": 3, "agent-supplychain": 3, "agent-trainingcost": 3, "agent-funding": 2, "agent-github-ai": 1, "agent-legal-intel": 1, "agent-regulations": 1, "agent-research": 1, "agent-mcp-registry": 1}, "surface": null}, {"key": "elevenlabs", "name": "elevenlabs.io", "domains": ["elevenlabs.io"], "mind_observations": 0, "dataset_count": 14, "row_mentions": 91, "datasets": {"agent-surface": 50, "agent-incidents": 14, "agent-funding": 4, "agent-tools": 4, "agent-models": 3, "agent-compensation": 2, "agent-regulations": 2, "agent-regulatoryfines": 2, "agent-research": 2, "agent-safety": 2, "agent-supplychain": 2, "agent-trainingcost": 2, "agent-cve": 1, "agent-github-ai": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "supabase", "name": "supabase.com", "domains": ["supabase.com"], "mind_observations": 0, "dataset_count": 13, "row_mentions": 150, "datasets": {"agent-incidents": 75, "agent-surface": 50, "agent-mcp-registry": 12, "agent-models": 3, "agent-stars": 2, "agent-compensation": 1, "agent-opensource": 1, "agent-pricing": 1, "agent-regulations": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "ai and liability", "name": "AI and Liability", "domains": [], "mind_observations": 4, "dataset_count": 13, "row_mentions": 107, "datasets": {"agent-incidents": 61, "agent-legal-intel": 15, "agent-regulations": 15, "agent-compensation": 2, "agent-regulatoryfines": 2, "agent-safety": 2, "agent-supplychain": 2, "agent-tools": 2, "agent-trainingcost": 2, "agent-acquisitions": 1, "agent-consumer": 1, "agent-funding": 1, "agent-models": 1}, "surface": null}, {"key": "spotify", "name": "spotify.com", "domains": ["spotify.com"], "mind_observations": 0, "dataset_count": 13, "row_mentions": 93, "datasets": {"agent-surface": 50, "agent-incidents": 27, "agent-models": 3, "agent-acquisitions": 2, "agent-regulations": 2, "agent-mcp-registry": 2, "agent-compensation": 1, "agent-funding": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-tools": 1, "agent-trainingcost": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "perplexity", "name": "perplexity.ai", "domains": ["perplexity.ai"], "mind_observations": 0, "dataset_count": 12, "row_mentions": 189, "datasets": {"agent-incidents": 79, "agent-surface": 50, "agent-research": 30, "agent-mcp-registry": 9, "agent-models": 6, "agent-tools": 6, "agent-domains": 2, "agent-regulations": 2, "agent-salary-live": 2, "agent-domain-expiry": 1, "agent-exec-moves": 1, "agent-legal-intel": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "show hn skills mcp", "name": "Show HN: Skills MCP", "domains": [], "mind_observations": 2, "dataset_count": 12, "row_mentions": 85, "datasets": {"agent-incidents": 57, "agent-models": 8, "agent-tools": 7, "agent-regulations": 3, "agent-failures": 2, "agent-funding": 2, "agent-compensation": 1, "agent-enterprise": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "midjourney", "name": "midjourney", "domains": [], "mind_observations": 4, "dataset_count": 12, "row_mentions": 19, "datasets": {"agent-incidents": 7, "agent-tools": 2, "agent-compensation": 1, "agent-domain-expiry": 1, "agent-exec-moves": 1, "agent-launches": 1, "agent-regulations": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-techstack": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "stripe", "name": "stripe.com", "domains": ["stripe.com"], "mind_observations": 0, "dataset_count": 11, "row_mentions": 246, "datasets": {"agent-incidents": 107, "agent-surface": 50, "agent-mcp-registry": 45, "agent-acquisitions": 13, "agent-funding": 9, "agent-regulations": 7, "agent-tools": 7, "agent-models": 5, "agent-canary": 1, "agent-research": 1, "agent-standards": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "modal", "name": "modal", "domains": [], "mind_observations": 132, "dataset_count": 11, "row_mentions": 221, "datasets": {"agent-compute-prices": 104, "agent-compute-arb": 70, "agent-incidents": 30, "agent-research": 8, "agent-models": 2, "agent-vanish": 2, "agent-domain-expiry": 1, "agent-funding": 1, "agent-inference": 1, "agent-regulations": 1, "agent-tools": 1}, "surface": null}, {"key": "zapier", "name": "zapier.com", "domains": ["zapier.com"], "mind_observations": 0, "dataset_count": 11, "row_mentions": 75, "datasets": {"agent-surface": 50, "agent-incidents": 16, "agent-compensation": 1, "agent-models": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-stars": 1, "agent-supplychain": 1, "agent-tools": 1, "agent-trainingcost": 1, "agent-mcp-registry": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "gemini 3 8 flash", "name": "Gemini 3.8 Flash", "domains": [], "mind_observations": 3, "dataset_count": 11, "row_mentions": 64, "datasets": {"agent-policydrift": 29, "agent-models": 19, "agent-incidents": 4, "agent-research": 3, "agent-regulations": 2, "agent-tools": 2, "agent-funding": 1, "agent-pricing": 1, "agent-releases": 1, "agent-sentiment": 1, "agent-stars": 1}, "surface": null}, {"key": "palantir", "name": "Palantir", "domains": [], "mind_observations": 13, "dataset_count": 11, "row_mentions": 44, "datasets": {"agent-incidents": 16, "agent-compensation": 4, "agent-regulatoryfines": 4, "agent-safety": 4, "agent-supplychain": 4, "agent-trainingcost": 4, "agent-procurement": 3, "agent-regulations": 2, "agent-funding": 1, "agent-partnerships": 1, "agent-sovereign": 1}, "surface": null}, {"key": "replicate", "name": "replicate", "domains": ["replicate.com"], "mind_observations": 132, "dataset_count": 10, "row_mentions": 330, "datasets": {"agent-compute-prices": 104, "agent-compute-arb": 68, "agent-policydrift": 57, "agent-surface": 50, "agent-incidents": 34, "agent-research": 12, "agent-domains": 2, "agent-acquisitions": 1, "agent-domain-expiry": 1, "agent-failures": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "you are the harness", "name": "You Are the Harness", "domains": [], "mind_observations": 3, "dataset_count": 10, "row_mentions": 145, "datasets": {"agent-incidents": 71, "agent-models": 25, "agent-tools": 16, "agent-funding": 10, "agent-regulations": 8, "agent-opensource": 4, "agent-pricing": 4, "agent-acquisitions": 3, "agent-failures": 3, "agent-supplychain": 1}, "surface": null}, {"key": "walmart", "name": "walmart.com", "domains": ["walmart.com"], "mind_observations": 0, "dataset_count": 10, "row_mentions": 68, "datasets": {"agent-surface": 50, "agent-incidents": 8, "agent-regulations": 2, "agent-tools": 2, "agent-consumer": 1, "agent-funding": 1, "agent-models": 1, "agent-pricing": 1, "agent-talent": 1, "agent-synthetic-data": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "fireworks", "name": "fireworks.ai", "domains": ["fireworks.ai"], "mind_observations": 0, "dataset_count": 10, "row_mentions": 64, "datasets": {"agent-surface": 50, "agent-incidents": 3, "agent-releases": 3, "agent-funding": 2, "agent-domain-expiry": 1, "agent-legal-intel": 1, "agent-models": 1, "agent-tools": 1, "agent-vanish": 1, "agent-orphan-adoption": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "xai", "name": "xai", "domains": [], "mind_observations": 43, "dataset_count": 10, "row_mentions": 55, "datasets": {"agent-incidents": 38, "agent-research": 5, "agent-regulations": 3, "agent-funding": 2, "agent-salary-live": 2, "agent-acquisitions": 1, "agent-chips": 1, "agent-exec-moves": 1, "agent-models": 1, "agent-stars": 1}, "surface": null}, {"key": "bytedance", "name": "bytedance", "domains": [], "mind_observations": 16, "dataset_count": 10, "row_mentions": 16, "datasets": {"agent-incidents": 3, "agent-chips": 2, "agent-funding": 2, "agent-legal-intel": 2, "agent-models": 2, "agent-acquisitions": 1, "agent-consumer": 1, "agent-regulations": 1, "agent-stars": 1, "agent-tools": 1}, "surface": null}, {"key": "swe-bench", "name": "SWE-Bench", "domains": [], "mind_observations": 12, "dataset_count": 9, "row_mentions": 149, "datasets": {"agent-benchmarks-v2": 115, "agent-research": 18, "agent-incidents": 7, "agent-models": 4, "agent-compensation": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "booking", "name": "booking.com", "domains": ["booking.com"], "mind_observations": 0, "dataset_count": 9, "row_mentions": 110, "datasets": {"agent-surface": 50, "agent-incidents": 34, "agent-mcp-registry": 14, "agent-regulations": 4, "agent-funding": 2, "agent-research": 2, "agent-tools": 2, "agent-energy-prices": 1, "agent-models": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "talk like claude day", "name": "Talk Like Claude Day", "domains": [], "mind_observations": 6, "dataset_count": 9, "row_mentions": 12, "datasets": {"agent-incidents": 3, "agent-models": 2, "agent-chips": 1, "agent-compensation": 1, "agent-opensource": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "ebay", "name": "ebay.com", "domains": ["ebay.com"], "mind_observations": 0, "dataset_count": 8, "row_mentions": 72, "datasets": {"agent-surface": 50, "agent-incidents": 14, "agent-regulations": 2, "agent-tools": 2, "agent-acquisitions": 1, "agent-research": 1, "agent-talent": 1, "agent-mcp-registry": 1}, "surface": {"score": 20, "llms_txt": false, "agent_card": false, "blocks": ["GPTBot", "ClaudeBot", "anthropic-ai", "PerplexityBot", "CCBot", "Bytespider", "Applebot-Extended", "meta-externalagent"]}}, {"key": "samsung", "name": "samsung", "domains": [], "mind_observations": 58, "dataset_count": 8, "row_mentions": 38, "datasets": {"agent-incidents": 21, "agent-funding": 6, "agent-regulations": 4, "agent-tools": 3, "agent-acquisitions": 1, "agent-inference": 1, "agent-models": 1, "agent-synthetic-data": 1}, "surface": null}, {"key": "paypal", "name": "paypal.com", "domains": ["paypal.com"], "mind_observations": 0, "dataset_count": 7, "row_mentions": 68, "datasets": {"agent-surface": 50, "agent-incidents": 12, "agent-funding": 2, "agent-acquisitions": 1, "agent-controversies": 1, "agent-failures": 1, "agent-models": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "airbnb", "name": "airbnb.com", "domains": ["airbnb.com"], "mind_observations": 0, "dataset_count": 7, "row_mentions": 66, "datasets": {"agent-surface": 50, "agent-incidents": 10, "agent-acquisitions": 2, "agent-executive-moves": 1, "agent-models": 1, "agent-partnerships": 1, "agent-tools": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": ["GPTBot", "ClaudeBot", "Applebot-Extended"]}}, {"key": "clearview", "name": "clearview", "domains": [], "mind_observations": 4, "dataset_count": 7, "row_mentions": 10, "datasets": {"agent-incidents": 3, "agent-reg-fines": 2, "agent-compensation": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "baidu", "name": "baidu", "domains": [], "mind_observations": 10, "dataset_count": 7, "row_mentions": 7, "datasets": {"agent-compensation": 1, "agent-incidents": 1, "agent-models": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "coreweave", "name": "coreweave", "domains": [], "mind_observations": 165, "dataset_count": 6, "row_mentions": 251, "datasets": {"agent-compute-prices": 130, "agent-compute-arb": 84, "agent-policydrift": 28, "agent-funding": 6, "agent-incidents": 2, "agent-domain-expiry": 1}, "surface": null}, {"key": "hubspot", "name": "hubspot.com", "domains": ["hubspot.com"], "mind_observations": 0, "dataset_count": 6, "row_mentions": 68, "datasets": {"agent-surface": 50, "agent-incidents": 11, "agent-mcp-registry": 3, "agent-regulations": 2, "agent-models": 1, "agent-tools": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "etsy", "name": "etsy.com", "domains": ["etsy.com"], "mind_observations": 0, "dataset_count": 6, "row_mentions": 59, "datasets": {"agent-surface": 50, "agent-incidents": 3, "agent-tools": 3, "agent-models": 1, "agent-regulations": 1, "agent-talent": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "claude public artifacts", "name": "Claude Public Artifacts", "domains": [], "mind_observations": 2, "dataset_count": 6, "row_mentions": 10, "datasets": {"agent-incidents": 3, "agent-models": 3, "agent-failures": 1, "agent-regulations": 1, "agent-research": 1, "agent-tools": 1}, "surface": null}, {"key": "mmlu", "name": "MMLU", "domains": [], "mind_observations": 12, "dataset_count": 5, "row_mentions": 157, "datasets": {"agent-benchmarks-v2": 138, "agent-research": 6, "agent-incidents": 5, "agent-models": 5, "agent-benchmarks": 3}, "surface": null}, {"key": "humaneval", "name": "HumanEval", "domains": [], "mind_observations": 12, "dataset_count": 5, "row_mentions": 155, "datasets": {"agent-benchmarks-v2": 138, "agent-research": 9, "agent-models": 4, "agent-benchmarks": 3, "agent-incidents": 1}, "surface": null}, {"key": "coinbase", "name": "coinbase.com", "domains": ["coinbase.com"], "mind_observations": 0, "dataset_count": 5, "row_mentions": 102, "datasets": {"agent-surface": 50, "agent-protocol-watch": 35, "agent-incidents": 14, "agent-acquisitions": 2, "agent-mcp-registry": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "llamaindex", "name": "llamaindex.ai", "domains": ["llamaindex.ai"], "mind_observations": 0, "dataset_count": 5, "row_mentions": 71, "datasets": {"agent-surface": 50, "agent-incidents": 18, "agent-cve": 1, "agent-models": 1, "agent-regulations": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "netflix", "name": "netflix.com", "domains": ["netflix.com"], "mind_observations": 0, "dataset_count": 5, "row_mentions": 70, "datasets": {"agent-surface": 50, "agent-incidents": 16, "agent-acquisitions": 2, "agent-funding": 1, "agent-models": 1}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "atlassian", "name": "atlassian.com", "domains": ["atlassian.com"], "mind_observations": 0, "dataset_count": 5, "row_mentions": 57, "datasets": {"agent-surface": 50, "agent-incidents": 4, "agent-enterprise": 1, "agent-funding": 1, "agent-tools": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "lemonde", "name": "lemonde.fr", "domains": ["lemonde.fr"], "mind_observations": 0, "dataset_count": 5, "row_mentions": 55, "datasets": {"agent-surface": 50, "agent-incidents": 2, "agent-models": 1, "agent-regulations": 1, "agent-sentiment": 1}, "surface": {"score": 20, "llms_txt": false, "agent_card": false, "blocks": ["ClaudeBot", "Claude-User", "anthropic-ai", "Google-Extended", "CCBot", "Bytespider", "Applebot-Extended"]}}, {"key": "databricks", "name": "databricks", "domains": [], "mind_observations": 9, "dataset_count": 5, "row_mentions": 14, "datasets": {"agent-incidents": 9, "agent-funding": 2, "agent-regulations": 1, "agent-research": 1, "agent-valuations": 1}, "surface": null}, {"key": "ea safety", "name": "EA Safety", "domains": [], "mind_observations": 5, "dataset_count": 5, "row_mentions": 5, "datasets": {"agent-compensation": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "the safety belt edition", "name": "The Safety Belt Edition", "domains": [], "mind_observations": 5, "dataset_count": 5, "row_mentions": 5, "datasets": {"agent-compensation": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "cannabis supply chain database", "name": "Cannabis Supply Chain Database", "domains": [], "mind_observations": 5, "dataset_count": 5, "row_mentions": 5, "datasets": {"agent-compensation": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "the cost of safetyism", "name": "The Cost of Safetyism", "domains": [], "mind_observations": 5, "dataset_count": 5, "row_mentions": 5, "datasets": {"agent-compensation": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "are refurbished phones dangerous", "name": "Are Refurbished Phones Dangerous?", "domains": [], "mind_observations": 5, "dataset_count": 5, "row_mentions": 5, "datasets": {"agent-compensation": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "harvey tenet research preview", "name": "Harvey Tenet Research Preview", "domains": [], "mind_observations": 5, "dataset_count": 5, "row_mentions": 5, "datasets": {"agent-compensation": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "compensation bands and promotions", "name": "Compensation Bands and Promotions", "domains": [], "mind_observations": 5, "dataset_count": 5, "row_mentions": 5, "datasets": {"agent-compensation": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "ai safety conspiracy theories", "name": "AI Safety Conspiracy Theories", "domains": [], "mind_observations": 5, "dataset_count": 5, "row_mentions": 5, "datasets": {"agent-compensation": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-supplychain": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "netlify", "name": "netlify", "domains": [], "mind_observations": 173, "dataset_count": 4, "row_mentions": 709, "datasets": {"agent-outages": 686, "agent-incidents": 13, "agent-mcp-registry": 9, "agent-funding": 1}, "surface": null}, {"key": "runpod", "name": "runpod", "domains": [], "mind_observations": 238, "dataset_count": 4, "row_mentions": 407, "datasets": {"agent-compute-prices": 278, "agent-compute-arb": 120, "agent-energy-futures": 7, "agent-compute": 2}, "surface": null}, {"key": "gpqa", "name": "GPQA", "domains": [], "mind_observations": 12, "dataset_count": 4, "row_mentions": 124, "datasets": {"agent-benchmarks-v2": 115, "agent-models": 7, "agent-research": 1, "agent-valuations": 1}, "surface": null}, {"key": "arc-challenge", "name": "ARC-Challenge", "domains": [], "mind_observations": 12, "dataset_count": 4, "row_mentions": 118, "datasets": {"agent-benchmarks-v2": 115, "agent-incidents": 1, "agent-models": 1, "agent-research": 1}, "surface": null}, {"key": "weaviate", "name": "weaviate.io", "domains": ["weaviate.io"], "mind_observations": 0, "dataset_count": 4, "row_mentions": 62, "datasets": {"agent-surface": 50, "agent-incidents": 9, "agent-models": 2, "agent-regulations": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "deepgram", "name": "deepgram.com", "domains": ["deepgram.com"], "mind_observations": 0, "dataset_count": 4, "row_mentions": 59, "datasets": {"agent-surface": 50, "agent-incidents": 7, "agent-inference": 1, "agent-models": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "safety warning carborane mixtures", "name": "Safety Warning: Carborane Mixtures", "domains": [], "mind_observations": 4, "dataset_count": 4, "row_mentions": 4, "datasets": {"agent-compensation": 1, "agent-regulatoryfines": 1, "agent-safety": 1, "agent-trainingcost": 1}, "surface": null}, {"key": "lambdalabs", "name": "lambdalabs", "domains": [], "mind_observations": 238, "dataset_count": 3, "row_mentions": 305, "datasets": {"agent-compute-prices": 182, "agent-compute-arb": 116, "agent-energy-futures": 7}, "surface": null}, {"key": "vastai avg", "name": "vastai_avg", "domains": [], "mind_observations": 201, "dataset_count": 3, "row_mentions": 267, "datasets": {"agent-compute-prices": 156, "agent-compute-arb": 108, "agent-energy-futures": 3}, "surface": null}, {"key": "expedia", "name": "expedia.com", "domains": ["expedia.com"], "mind_observations": 0, "dataset_count": 3, "row_mentions": 52, "datasets": {"agent-surface": 50, "agent-incidents": 1, "agent-partnerships": 1}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "show hn time wizard", "name": "Show HN: Time Wizard", "domains": [], "mind_observations": 2, "dataset_count": 3, "row_mentions": 7, "datasets": {"agent-incidents": 5, "agent-regulations": 1, "agent-tools": 1}, "surface": null}, {"key": "programbench", "name": "ProgramBench", "domains": [], "mind_observations": 2, "dataset_count": 3, "row_mentions": 7, "datasets": {"agent-research": 4, "agent-models": 2, "agent-regulations": 1}, "surface": null}, {"key": "the defender s window", "name": "The Defender’s Window", "domains": [], "mind_observations": 2, "dataset_count": 3, "row_mentions": 4, "datasets": {"agent-incidents": 2, "agent-models": 1, "agent-research": 1}, "surface": null}, {"key": "gemini 3 5 transcribe", "name": "Gemini 3.5 Transcribe", "domains": [], "mind_observations": 2, "dataset_count": 3, "row_mentions": 3, "datasets": {"agent-incidents": 1, "agent-models": 1, "agent-tools": 1}, "surface": null}, {"key": "omarchy histerya", "name": "Omarchy Histerya", "domains": [], "mind_observations": 2, "dataset_count": 3, "row_mentions": 3, "datasets": {"agent-incidents": 1, "agent-models": 1, "agent-regulations": 1}, "surface": null}, {"key": "heroku", "name": "heroku", "domains": [], "mind_observations": 173, "dataset_count": 2, "row_mentions": 690, "datasets": {"agent-outages": 686, "agent-incidents": 4}, "surface": null}, {"key": "hellaswag", "name": "HellaSwag", "domains": [], "mind_observations": 12, "dataset_count": 2, "row_mentions": 93, "datasets": {"agent-benchmarks-v2": 92, "agent-incidents": 1}, "surface": null}, {"key": "linode", "name": "linode", "domains": [], "mind_observations": 75, "dataset_count": 2, "row_mentions": 76, "datasets": {"agent-vps-prices": 75, "agent-incidents": 1}, "surface": null}, {"key": "pinecone", "name": "pinecone.io", "domains": ["pinecone.io"], "mind_observations": 0, "dataset_count": 2, "row_mentions": 71, "datasets": {"agent-surface": 50, "agent-incidents": 21}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "stackoverflow", "name": "stackoverflow.com", "domains": ["stackoverflow.com"], "mind_observations": 0, "dataset_count": 2, "row_mentions": 59, "datasets": {"agent-surface": 50, "agent-incidents": 9}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "klarna", "name": "klarna.com", "domains": ["klarna.com"], "mind_observations": 0, "dataset_count": 2, "row_mentions": 54, "datasets": {"agent-surface": 50, "agent-mcp-registry": 4}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "invisible agency", "name": "Invisible Agency", "domains": [], "mind_observations": 2, "dataset_count": 2, "row_mentions": 4, "datasets": {"agent-incidents": 2, "agent-jobs": 2}, "surface": null}, {"key": "mistral raises 3b", "name": "Mistral raises €3B", "domains": [], "mind_observations": 2, "dataset_count": 2, "row_mentions": 3, "datasets": {"agent-funding": 2, "agent-models": 1}, "surface": null}, {"key": "spatially adaptive noise injection", "name": "Spatially Adaptive Noise Injection", "domains": [], "mind_observations": 2, "dataset_count": 2, "row_mentions": 2, "datasets": {"agent-arxiv": 1, "agent-research": 1}, "surface": null}, {"key": "the cost of seamlessness", "name": "The Cost of Seamlessness", "domains": [], "mind_observations": 2, "dataset_count": 2, "row_mentions": 2, "datasets": {"agent-funding": 1, "agent-incidents": 1}, "surface": null}, {"key": "asl-0 license", "name": "ASL-0 License", "domains": [], "mind_observations": 2, "dataset_count": 2, "row_mentions": 2, "datasets": {"agent-incidents": 1, "agent-regulations": 1}, "surface": null}, {"key": "mastra agent factory", "name": "Mastra Agent Factory", "domains": [], "mind_observations": 2, "dataset_count": 2, "row_mentions": 2, "datasets": {"agent-models": 1, "agent-regulations": 1}, "surface": null}, {"key": "gimlet s series b", "name": "Gimlet's Series B", "domains": [], "mind_observations": 2, "dataset_count": 2, "row_mentions": 2, "datasets": {"agent-funding": 1, "agent-regulations": 1}, "surface": null}, {"key": "testing ads in chatgpt", "name": "Testing ads in ChatGPT", "domains": [], "mind_observations": 2, "dataset_count": 2, "row_mentions": 2, "datasets": {"agent-incidents": 1, "agent-models": 1}, "surface": null}, {"key": "openai joins ports-pike project", "name": "OpenAI joins PORTS-Pike project", "domains": [], "mind_observations": 2, "dataset_count": 2, "row_mentions": 2, "datasets": {"agent-incidents": 1, "agent-models": 1}, "surface": null}, {"key": "scale-ai", "name": "scale-ai", "domains": [], "mind_observations": 245, "dataset_count": 1, "row_mentions": 253, "datasets": {"agent-jobs": 253}, "surface": null}, {"key": "lmsys-arena", "name": "LMSYS-Arena", "domains": [], "mind_observations": 12, "dataset_count": 1, "row_mentions": 138, "datasets": {"agent-benchmarks-v2": 138}, "surface": null}, {"key": "adyen", "name": "adyen.com", "domains": ["adyen.com"], "mind_observations": 0, "dataset_count": 1, "row_mentions": 50, "datasets": {"agent-surface": 50}, "surface": {"score": 65, "llms_txt": true, "agent_card": false, "blocks": []}}, {"key": "zalando", "name": "zalando.com", "domains": ["zalando.com"], "mind_observations": 0, "dataset_count": 1, "row_mentions": 50, "datasets": {"agent-surface": 50}, "surface": {"score": 30, "llms_txt": false, "agent_card": false, "blocks": []}}, {"key": "vultr", "name": "vultr", "domains": [], "mind_observations": 12, "dataset_count": 1, "row_mentions": 12, "datasets": {"agent-vps-prices": 12}, "surface": null}, {"key": "imerit technology", "name": "iMerit Technology", "domains": [], "mind_observations": 2, "dataset_count": 1, "row_mentions": 2, "datasets": {"agent-jobs": 2}, "surface": null}, {"key": "widgo", "name": "Widgo", "domains": [], "mind_observations": 2, "dataset_count": 1, "row_mentions": 2, "datasets": {"agent-tools": 2}, "surface": null}], "negative_space": [{"subject": "openai", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 71, "mcp_rows": 9}}, {"subject": "anthropic", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 65, "mcp_rows": 1}}, {"subject": "google", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 58, "top": ["agent-incidents", "agent-models", "agent-benchmarks-v2", "agent-temp0", "agent-regulations"]}}, {"subject": "github", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 46, "domains": ["github.com"]}}, {"subject": "github", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 46, "mcp_rows": 1489}}, {"subject": "meta", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 44, "top": ["agent-incidents", "agent-benchmarks-v2", "agent-sentiment", "agent-regulations", "agent-tools"]}}, {"subject": "together", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 40, "domains": ["together.ai"]}}, {"subject": "together", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 40, "mcp_rows": 3}}, {"subject": "medium", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 39, "mcp_rows": 2}}, {"subject": "microsoft", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 35, "top": ["agent-incidents", "agent-models", "agent-acquisitions", "agent-tools", "agent-funding"]}}, {"subject": "amazon", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 34, "mcp_rows": 17}}, {"subject": "apple", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 33, "top": ["agent-incidents", "agent-models", "agent-tools", "agent-releases", "agent-mcp-registry"]}}, {"subject": "mistral", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 32, "domains": ["mistral.ai"]}}, {"subject": "mistral", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 32, "mcp_rows": 1}}, {"subject": "nvidia", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 31, "top": ["agent-incidents", "agent-funding", "agent-regulations", "agent-research", "agent-models"]}}, {"subject": "math", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 27, "top": ["agent-research", "agent-incidents", "agent-benchmarks-v2", "agent-temp0", "agent-models"]}}, {"subject": "deepseek", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 27, "mcp_rows": 2}}, {"subject": "target", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 25, "domains": ["target.com"]}}, {"subject": "target", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 25, "mcp_rows": 1}}, {"subject": "reddit", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 25, "mcp_rows": 13}}, {"subject": "wikipedia", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 25, "mcp_rows": 4}}, {"subject": "groq", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 24, "mcp_rows": 1}}, {"subject": "hi hn", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 24, "top": ["agent-incidents", "agent-tools", "agent-models", "agent-regulations", "agent-funding"]}}, {"subject": "vercel", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 23, "domains": ["vercel.com"]}}, {"subject": "vercel", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 23, "mcp_rows": 18}}, {"subject": "stability", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 23, "top": ["agent-research", "agent-incidents", "agent-trials", "agent-regulations", "agent-retractions"]}}, {"subject": "stable", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 22, "top": ["agent-research", "agent-incidents", "agent-oracle", "agent-models", "agent-releases"]}}, {"subject": "ollama", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 22, "top": ["agent-incidents", "agent-models", "agent-drift", "agent-cve", "agent-canary"]}}, {"subject": "beta", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 22, "top": ["agent-incidents", "agent-models", "agent-tools", "agent-releases", "agent-regulations"]}}, {"subject": "huggingface", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 21, "mcp_rows": 4}}, {"subject": "langchain", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 21, "mcp_rows": 14}}, {"subject": "n8n", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 21, "domains": ["n8n.io"]}}, {"subject": "n8n", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 21, "mcp_rows": 3}}, {"subject": "linear", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 20, "domains": ["linear.app"]}}, {"subject": "linear", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 20, "mcp_rows": 3}}, {"subject": "amd", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 20, "top": ["agent-models", "agent-incidents", "agent-regulations", "agent-acquisitions", "agent-funding"]}}, {"subject": "notion", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 19, "domains": ["notion.so"]}}, {"subject": "notion", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 19, "mcp_rows": 2}}, {"subject": "cohere", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 16, "domains": ["cohere.com"]}}, {"subject": "intel", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 16, "top": ["agent-incidents", "agent-models", "agent-research", "agent-jobs", "agent-mcp-registry"]}}, {"subject": "shopify", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 15, "domains": ["shopify.com"]}}, {"subject": "shopify", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 15, "mcp_rows": 10}}, {"subject": "salesforce", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 15, "domains": ["salesforce.com"]}}, {"subject": "salesforce", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 15, "mcp_rows": 8}}, {"subject": "tesla", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 15, "top": ["agent-incidents", "agent-energy-prices", "agent-compute-prices", "agent-compensation", "agent-models"]}}, {"subject": "elevenlabs", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 14, "domains": ["elevenlabs.io"]}}, {"subject": "supabase", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 13, "domains": ["supabase.com"]}}, {"subject": "supabase", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 13, "mcp_rows": 12}}, {"subject": "ai and liability", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 13, "top": ["agent-incidents", "agent-legal-intel", "agent-regulations", "agent-compensation", "agent-regulatoryfines"]}}, {"subject": "spotify", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 13, "mcp_rows": 2}}, {"subject": "perplexity", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 12, "mcp_rows": 9}}, {"subject": "show hn skills mcp", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 12, "top": ["agent-incidents", "agent-models", "agent-tools", "agent-regulations", "agent-failures"]}}, {"subject": "midjourney", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 12, "top": ["agent-incidents", "agent-tools", "agent-compensation", "agent-domain-expiry", "agent-exec-moves"]}}, {"subject": "stripe", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 11, "domains": ["stripe.com"]}}, {"subject": "stripe", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 11, "mcp_rows": 45}}, {"subject": "modal", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 11, "top": ["agent-compute-prices", "agent-compute-arb", "agent-incidents", "agent-research", "agent-models"]}}, {"subject": "zapier", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 11, "domains": ["zapier.com"]}}, {"subject": "zapier", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 11, "mcp_rows": 1}}, {"subject": "gemini 3 8 flash", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 11, "top": ["agent-policydrift", "agent-models", "agent-incidents", "agent-research", "agent-regulations"]}}, {"subject": "palantir", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 11, "top": ["agent-incidents", "agent-compensation", "agent-regulatoryfines", "agent-safety", "agent-supplychain"]}}, {"subject": "replicate", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 10, "domains": ["replicate.com"]}}, {"subject": "you are the harness", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 10, "top": ["agent-incidents", "agent-models", "agent-tools", "agent-funding", "agent-regulations"]}}, {"subject": "xai", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 10, "top": ["agent-incidents", "agent-research", "agent-regulations", "agent-funding", "agent-salary-live"]}}, {"subject": "bytedance", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 10, "top": ["agent-incidents", "agent-chips", "agent-funding", "agent-legal-intel", "agent-models"]}}, {"subject": "swe-bench", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 9, "top": ["agent-benchmarks-v2", "agent-research", "agent-incidents", "agent-models", "agent-compensation"]}}, {"subject": "booking", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 9, "mcp_rows": 14}}, {"subject": "talk like claude day", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 9, "top": ["agent-incidents", "agent-models", "agent-chips", "agent-compensation", "agent-opensource"]}}, {"subject": "ebay", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 8, "mcp_rows": 1}}, {"subject": "samsung", "gap": "high_signal_entity_not_in_surface_scan", "evidence": {"dataset_count": 8, "top": ["agent-incidents", "agent-funding", "agent-regulations", "agent-tools", "agent-acquisitions"]}}, {"subject": "paypal", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 7, "domains": ["paypal.com"]}}, {"subject": "hubspot", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 6, "domains": ["hubspot.com"]}}, {"subject": "hubspot", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 6, "mcp_rows": 3}}, {"subject": "etsy", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 6, "domains": ["etsy.com"]}}, {"subject": "coinbase", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 5, "domains": ["coinbase.com"]}}, {"subject": "coinbase", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 5, "mcp_rows": 1}}, {"subject": "atlassian", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 5, "domains": ["atlassian.com"]}}, {"subject": "weaviate", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 4, "domains": ["weaviate.io"]}}, {"subject": "deepgram", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 4, "domains": ["deepgram.com"]}}, {"subject": "expedia", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 3, "domains": ["expedia.com"]}}, {"subject": "klarna", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 2, "domains": ["klarna.com"]}}, {"subject": "klarna", "gap": "mcp_servers_listed_without_agent_card", "evidence": {"dataset_count": 2, "mcp_rows": 4}}, {"subject": "adyen", "gap": "llms_txt_without_agent_card", "evidence": {"dataset_count": 1, "domains": ["adyen.com"]}}], "summary": {"mcp_servers_listed_without_agent_card": 30, "high_signal_entity_not_in_surface_scan": 26, "llms_txt_without_agent_card": 26}}, "provenance": ["/home/legion33/legion-mind/legion_mind.db", "/home/legion33/agent-surface/surface.db", "/mnt/legion/agents/agent-acquisitions/acquisitions.db", "/mnt/legion/agents/agent-api-pricing/pricing.db", "/mnt/legion/agents/agent-apistatus/apistatus.db", "/mnt/legion/agents/agent-arxiv/papers.db", "/mnt/legion/agents/agent-attest/attest.db", "/mnt/legion/agents/agent-benchmarks/benchmarks.db", "/mnt/legion/agents/agent-benchmarks-v2/benchmarks.db", "/mnt/legion/agents/agent-canary/canary.db", "/mnt/legion/agents/agent-capabilities/capabilities.db", "/mnt/legion/agents/agent-changelog/changelog.db", "/mnt/legion/agents/agent-chips/chips.db", "/mnt/legion/agents/agent-climate-compute/climate.db", "/mnt/legion/agents/agent-compensation/compensation.db", "/mnt/legion/agents/agent-compute/compute.db", "/mnt/legion/agents/agent-compute-arb/arb.db", "/mnt/legion/agents/agent-compute-energy/compute_energy.db", "/mnt/legion/agents/agent-compute-prices/compute.db", "/mnt/legion/agents/agent-consumer/consumer.db", "/mnt/legion/agents/agent-contracts/contracts.db", "/mnt/legion/agents/agent-controversies/controversies.db", "/mnt/legion/agents/agent-crypto/crypto.db", "/mnt/legion/agents/agent-cve/cve.db", "/mnt/legion/agents/agent-cyprus/cyprus.db", "/mnt/legion/agents/agent-darkpool/darkpool.db", "/mnt/legion/agents/agent-datacenters/datacenters.db", "/mnt/legion/agents/agent-deployments/deployments.db", "/mnt/legion/agents/agent-domain-expiry/expiry.db", "/mnt/legion/agents/agent-domains/domains.db", "/mnt/legion/agents/agent-drift/drift.db", "/mnt/legion/agents/agent-earnings/earnings.db", "/mnt/legion/agents/agent-energy/energy.db", "/mnt/legion/agents/agent-energy-futures/energy.db", "/mnt/legion/agents/agent-energy-prices/energy.db", "/mnt/legion/agents/agent-enterprise/enterprise.db", "/mnt/legion/agents/agent-ethics/ethics.db", "/mnt/legion/agents/agent-exec-moves/exec.db", "/mnt/legion/agents/agent-executive-moves/moves.db", "/mnt/legion/agents/agent-failures/failures.db", "/mnt/legion/agents/agent-funding/funding.db", "/mnt/legion/agents/agent-ghost/ghost.db", "/mnt/legion/agents/agent-github-ai/repos.db", "/mnt/legion/agents/agent-grants/grants.db", "/mnt/legion/agents/agent-hf-trending/trending.db", "/mnt/legion/agents/agent-hnjobs/jobs.db", "/mnt/legion/agents/agent-incidents/incidents.db", "/mnt/legion/agents/agent-inference/inference.db", "/mnt/legion/agents/agent-insurance/insurance.db", "/mnt/legion/agents/agent-jobs/jobs.db", "/mnt/legion/agents/agent-latency/latency.db", "/mnt/legion/agents/agent-launches/launches.db", "/mnt/legion/agents/agent-lawsuits/lawsuits.db", "/mnt/legion/agents/agent-layoffs/layoffs.db", "/mnt/legion/agents/agent-leads/leads.db", "/mnt/legion/agents/agent-legal/legal.db", "/mnt/legion/agents/agent-legal-intel/legal.db", "/mnt/legion/agents/agent-licenses/licenses.db", "/mnt/legion/agents/agent-lobbying/lobbying.db", "/mnt/legion/agents/agent-media/media.db", "/mnt/legion/agents/agent-military/military.db", "/mnt/legion/agents/agent-models/models.db", "/mnt/legion/agents/agent-opensource/opensource.db", "/mnt/legion/agents/agent-oracle/oracle.db", "/mnt/legion/agents/agent-outages/outages.db", "/mnt/legion/agents/agent-papers/papers.db", "/mnt/legion/agents/agent-partnerships/partnerships.db", "/mnt/legion/agents/agent-policydrift/policydrift.db", "/mnt/legion/agents/agent-pricing/pricing.db", "/mnt/legion/agents/agent-pricing-v2/pricing.db", "/mnt/legion/agents/agent-procurement/procurement.db", "/mnt/legion/agents/agent-prompts/prompts.db", "/mnt/legion/agents/agent-quantum/quantum.db", "/mnt/legion/agents/agent-ratelimits/ratelimits.db", "/mnt/legion/agents/agent-realestate/realestate.db", "/mnt/legion/agents/agent-reddit-mentions/mentions.db", "/mnt/legion/agents/agent-reg-fines/fines.db", "/mnt/legion/agents/agent-regulations/regulations.db", "/mnt/legion/agents/agent-regulatoryfines/regulatoryfines.db", "/mnt/legion/agents/agent-releases/releases.db", "/mnt/legion/agents/agent-research/research.db", "/mnt/legion/agents/agent-safety/safety.db", "/mnt/legion/agents/agent-security/security.db", "/mnt/legion/agents/agent-sentiment/sentiment.db", "/mnt/legion/agents/agent-sovereign/sovereign.db", "/mnt/legion/agents/agent-space-compute/space.db", "/mnt/legion/agents/agent-standards/standards.db", "/mnt/legion/agents/agent-stars/stars.db", "/mnt/legion/agents/agent-supply/supply.db", "/mnt/legion/agents/agent-supplychain/supplychain.db", "/mnt/legion/agents/agent-talent/talent.db", "/mnt/legion/agents/agent-talent-flow/talent.db", "/mnt/legion/agents/agent-techstack/techstack.db", "/mnt/legion/agents/agent-tokenomics/tokenomics.db", "/mnt/legion/agents/agent-tools/tools.db", "/mnt/legion/agents/agent-trainingcost/trainingcost.db", "/mnt/legion/agents/agent-valuations/valuations.db", "/mnt/legion/agents/agent-vps-prices/vps.db", "/mnt/legion/agents/agent-water-nexus/water.db", "/mnt/legion/agents/agent-wayback-prices/prices.db", "/mnt/legion/agents/agent-neuromorphic/neuro.db", "/mnt/legion/agents/agent-lobbying-money/lobby.db", "/mnt/legion/agents/agent-salary-live/salary.db", "/mnt/legion/agents/agent-synthetic-data/synth.db", "/mnt/legion/agents/agent-carbon/carbon.db", "/mnt/legion/agents/agent-economy/economy.db", "/home/legion33/agent-mcp-registry/mcp.db", "/home/legion33/agent-protocol-watch/protocols.db", "/home/legion33/agent-temp0/temp0.db", "/home/legion33/agent-vanish/vanish.db", "/home/legion33/agent-trials/trials.db", "/home/legion33/agent-retractions/retractions.db", "/home/legion33/agent-taboo-market/taboo.db", "/home/legion33/agent-reclaim/reclaim.db", "/home/legion33/agent-surface-source/surface_source.db", "/home/legion33/agent-thresholds/thresholds.db", "/mnt/legion/agents/agent-orphan-adoption/adoption.db", "/mnt/legion/agents/agent-orbital-compute/orbital.db"], "confidence": 0.6}
+{
+ "criteria": "scores attribués à des entités réelles nommées, dérivés de bases internes, sans URL de source",
+ "file": "evidence/entities.json",
+ "original_bytes": 79025,
+ "original_sha256": "52be94dfef707193f006b1722e3f8ce88afb6d0a7be5cae710969c6ae0f12e04",
+ "reason": "fabricated/unsourced claims about third parties",
+ "referenced_by": [
+  "snapshot-entities.json"
+ ],
+ "retracted_at": "2026-10-12T13:28:36Z",
+ "retraction_note": "retractions/2026-10-12_evidence-retraction.json",
+ "schema": "phoenix-tombstone/0.1",
+ "signature": {
+  "alg": "Ed25519",
+  "public_key_b64": "ENTr9tUMHipzZIBmrau7pH7+Py6szwCaj/9HKRZp4P8=",
+  "sig": "oKPPzTwvumwcG8oxCyukU1BOEjst7YPxi/oBtNGUqPtMx1iSPILmOFc7xWSKKKujrMxACwIROAbLZAC0+L2LDQ==",
+  "signed_message": "PHOENIX_TOMBSTONE_V0.1|abe1d6ead982674e0e81dfa83ac0760164ec8134c403ace2a4338f8428a28fae"
+ },
+ "status": "RETRACTED"
+}

evidence/fragility_confidence.json

@@ -1,307 +1,20 @@
 {
-  "status": "ok",
-  "generated_at": "2026-09-25T14:25:29+00:00",
-  "method": "fragility annotated by probe basis; 405/auth-gated latency is not inference latency",
-  "results": [
-    {
-      "provider": "anthropic",
-      "fragility": 0.552,
-      "fragility_class": "brittle",
-      "vix_30d": 2.666,
-      "cascade_rate": 33.3,
-      "latency_avg_ms": 221.7,
-      "min_input_price": 0.25,
-      "max_context": 200000,
-      "eligibility": "cheap_batch,long_context",
-      "route_score": 2.587,
-      "route_verdict": "avoid",
-      "high_stakes_ok": 0,
-      "reasons": "brittle(vix=2.666,cascade=33.3) | measured_on_405_responses: vix calcule sur latences de reponses method-gated, pas sur inference reelle",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00",
-      "confidence": "low",
-      "confidence_reason": "computed on method-gated 405 probes; not inference latency",
-      "probe_basis": "method_gated"
-    },
-    {
-      "provider": "cohere",
-      "fragility": 0.298,
-      "fragility_class": "robust",
-      "vix_30d": 1.456,
-      "cascade_rate": 2.2,
-      "latency_avg_ms": 597.1,
-      "min_input_price": 0.15,
-      "max_context": 128000,
-      "eligibility": "cheap_batch,long_context",
-      "route_score": 1.298,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "aucun signal defavorable mesure",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00",
-      "confidence": "medium_low",
-      "confidence_reason": "computed on auth-gated probes; reachability not inference latency",
-      "probe_basis": "auth_gated"
-    },
-    {
-      "provider": "deepseek",
-      "fragility": null,
-      "fragility_class": "unmeasured",
-      "vix_30d": null,
-      "cascade_rate": null,
-      "latency_avg_ms": null,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 1.95,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing; fragility_unmeasured(0.3 par defaut)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00",
-      "confidence": "low",
-      "confidence_reason": "probe basis weak or unavailable",
-      "probe_basis": "weak_probe_basis"
-    },
-    {
-      "provider": "google",
-      "fragility": 0.384,
-      "fragility_class": "watch",
-      "vix_30d": 1.832,
-      "cascade_rate": 3.2,
-      "latency_avg_ms": 495.8,
-      "min_input_price": 0.0375,
-      "max_context": 2000000,
-      "eligibility": "cheap_batch,long_context",
-      "route_score": 0.346,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "aucun signal defavorable mesure",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00",
-      "confidence": "medium_low",
-      "confidence_reason": "computed on auth-gated probes; reachability not inference latency",
-      "probe_basis": "auth_gated"
-    },
-    {
-      "provider": "groq",
-      "fragility": 0.391,
-      "fragility_class": "watch",
-      "vix_30d": 1.456,
-      "cascade_rate": 33.3,
-      "latency_avg_ms": 828.2,
-      "min_input_price": 0.05,
-      "max_context": 131072,
-      "eligibility": "cheap_batch,long_context",
-      "route_score": 0.464,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "slow(828.2ms)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00",
-      "confidence": "medium",
-      "confidence_reason": "computed on partially usable public probes",
-      "probe_basis": "usable_or_mixed"
-    },
-    {
-      "provider": "huggingface",
-      "fragility": 0.322,
-      "fragility_class": "watch",
-      "vix_30d": 1.583,
-      "cascade_rate": 1.9,
-      "latency_avg_ms": 591.3,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 1.983,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00",
-      "confidence": "medium",
-      "confidence_reason": "computed on partially usable public probes",
-      "probe_basis": "usable_or_mixed"
-    },
-    {
-      "provider": "meta",
-      "fragility": null,
-      "fragility_class": "unmeasured",
-      "vix_30d": null,
-      "cascade_rate": null,
-      "latency_avg_ms": null,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 1.95,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing; fragility_unmeasured(0.3 par defaut)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00",
-      "confidence": "low",
-      "confidence_reason": "probe basis weak or unavailable",
-      "probe_basis": "weak_probe_basis"
-    },
-    {
-      "provider": "mistral",
-      "fragility": 0.408,
-      "fragility_class": "watch",
-      "vix_30d": 2.009,
-      "cascade_rate": 2.5,
-      "latency_avg_ms": 416.5,
-      "min_input_price": 0.15,
-      "max_context": 262144,
-      "eligibility": "cheap_batch,long_context",
-      "route_score": 1.408,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "aucun signal defavorable mesure",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00",
-      "confidence": "medium_low",
-      "confidence_reason": "computed on auth-gated probes; reachability not inference latency",
-      "probe_basis": "auth_gated"
-    },
-    {
-      "provider": "nvidia",
-      "fragility": null,
-      "fragility_class": "unmeasured",
-      "vix_30d": null,
-      "cascade_rate": null,
-      "latency_avg_ms": null,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 1.95,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing; fragility_unmeasured(0.3 par defaut)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00",
-      "confidence": "low",
-      "confidence_reason": "probe basis weak or unavailable",
-      "probe_basis": "weak_probe_basis"
-    },
-    {
-      "provider": "ollama_hub",
-      "fragility": 0.416,
-      "fragility_class": "watch",
-      "vix_30d": 1.579,
-      "cascade_rate": 33.3,
-      "latency_avg_ms": 678.9,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 2.124,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00",
-      "confidence": "medium",
-      "confidence_reason": "computed on partially usable public probes",
-      "probe_basis": "usable_or_mixed"
-    },
-    {
-      "provider": "openai",
-      "fragility": 0.392,
-      "fragility_class": "watch",
-      "vix_30d": 1.635,
-      "cascade_rate": 2.2,
-      "latency_avg_ms": 414.0,
-      "min_input_price": 0.15,
-      "max_context": 200000,
-      "eligibility": "cheap_batch,long_context",
-      "route_score": 1.392,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "aucun signal defavorable mesure",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00",
-      "confidence": "medium_low",
-      "confidence_reason": "computed on auth-gated probes; reachability not inference latency",
-      "probe_basis": "auth_gated"
-    },
-    {
-      "provider": "qwen",
-      "fragility": null,
-      "fragility_class": "unmeasured",
-      "vix_30d": null,
-      "cascade_rate": null,
-      "latency_avg_ms": null,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 1.95,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing; fragility_unmeasured(0.3 par defaut)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00",
-      "confidence": "low",
-      "confidence_reason": "probe basis weak or unavailable",
-      "probe_basis": "weak_probe_basis"
-    },
-    {
-      "provider": "together",
-      "fragility": 0.283,
-      "fragility_class": "robust",
-      "vix_30d": 1.248,
-      "cascade_rate": 11.1,
-      "latency_avg_ms": 884.5,
-      "min_input_price": 0.18,
-      "max_context": 131072,
-      "eligibility": "cheap_batch,long_context",
-      "route_score": 1.54,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "slow(884.5ms)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00",
-      "confidence": "medium_low",
-      "confidence_reason": "computed on auth-gated probes; reachability not inference latency",
-      "probe_basis": "auth_gated"
-    },
-    {
-      "provider": "xai",
-      "fragility": null,
-      "fragility_class": "unmeasured",
-      "vix_30d": null,
-      "cascade_rate": null,
-      "latency_avg_ms": null,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 1.95,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing; fragility_unmeasured(0.3 par defaut)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00",
-      "confidence": "low",
-      "confidence_reason": "probe basis weak or unavailable",
-      "probe_basis": "weak_probe_basis"
-    },
-    {
-      "provider": "zhipu",
-      "fragility": null,
-      "fragility_class": "unmeasured",
-      "vix_30d": null,
-      "cascade_rate": null,
-      "latency_avg_ms": null,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 1.95,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing; fragility_unmeasured(0.3 par defaut)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00",
-      "confidence": "low",
-      "confidence_reason": "probe basis weak or unavailable",
-      "probe_basis": "weak_probe_basis"
-    }
-  ]
-}
+ "criteria": "scores de fragilité/route attribués à 12 fournisseurs réels sans URL de source",
+ "file": "evidence/fragility_confidence.json",
+ "original_bytes": 11023,
+ "original_sha256": "dc0b6481fede7a3efac6cc3683f1e73c6c6f11210633eec94a17555d63dba565",
+ "reason": "fabricated/unsourced claims about third parties",
+ "referenced_by": [
+  "snapshot-fragility_confidence.json"
+ ],
+ "retracted_at": "2026-10-12T13:28:36Z",
+ "retraction_note": "retractions/2026-10-12_evidence-retraction.json",
+ "schema": "phoenix-tombstone/0.1",
+ "signature": {
+  "alg": "Ed25519",
+  "public_key_b64": "ENTr9tUMHipzZIBmrau7pH7+Py6szwCaj/9HKRZp4P8=",
+  "sig": "q6A3fRcGaBu4+8xJXh6MdZ+t0XdiZUBqtfmvzDJozQIvT43wCuoVwheQrV0JSwKnmLO8rL5Ua3pwR4mK5yrNDQ==",
+  "signed_message": "PHOENIX_TOMBSTONE_V0.1|33093c27f8ab8c3330269ffd77b7fbc741020af9403a7d65457c2514c74d1c4a"
+ },
+ "status": "RETRACTED"
+}

evidence/orphan_adoption.json

@@ -1,179 +1,20 @@
 {
-  "organ": "LEGION-ORPHAN-ADOPTION",
-  "updated": "2026-09-27T11:34:27.990621+00:00",
-  "doctrine": "fork = human gate; non-permissive license = clean-room successor only",
-  "orphans_detected": 150,
-  "enriched": 15,
-  "adopt_candidates": 15,
-  "successor_only": 0,
-  "specs_emitted": 3,
-  "radar": [
-    {
-      "repo": "nvbn/thefuck",
-      "stars": 97884,
-      "last_push": "2024-07-19",
-      "license": "MIT",
-      "archived": false,
-      "open_issues": 461,
-      "classification": "ADOPT_CANDIDATE",
-      "action": "fork_requires_human_gate",
-      "measured": true
-    },
-    {
-      "repo": "ryanmcdermott/clean-code-javascript",
-      "stars": 94751,
-      "last_push": "2024-07-29",
-      "license": "MIT",
-      "archived": false,
-      "open_issues": 123,
-      "classification": "ADOPT_CANDIDATE",
-      "action": "fork_requires_human_gate",
-      "measured": true
-    },
-    {
-      "repo": "d2l-ai/d2l-zh",
-      "stars": 81116,
-      "last_push": "2024-07-30",
-      "license": "Apache-2.0",
-      "archived": false,
-      "open_issues": 127,
-      "classification": "ADOPT_CANDIDATE",
-      "action": "fork_requires_human_gate",
-      "measured": true
-    },
-    {
-      "repo": "kelseyhightower/nocode",
-      "stars": 65816,
-      "last_push": "2024-08-07",
-      "license": "Apache-2.0",
-      "archived": false,
-      "open_issues": 4721,
-      "classification": "ADOPT_CANDIDATE",
-      "action": "fork_requires_human_gate",
-      "measured": true
-    },
-    {
-      "repo": "lydiahallie/javascript-questions",
-      "stars": 65303,
-      "last_push": "2024-08-04",
-      "license": "MIT",
-      "archived": false,
-      "open_issues": 53,
-      "classification": "ADOPT_CANDIDATE",
-      "action": "fork_requires_human_gate",
-      "measured": true
-    },
-    {
-      "repo": "FiloSottile/mkcert",
-      "stars": 59695,
-      "last_push": "2024-08-13",
-      "license": "BSD-3-Clause",
-      "archived": false,
-      "open_issues": 177,
-      "classification": "ADOPT_CANDIDATE",
-      "action": "fork_requires_human_gate",
-      "measured": true
-    },
-    {
-      "repo": "tiimgreen/github-cheat-sheet",
-      "stars": 59278,
-      "last_push": "2024-04-15",
-      "license": "MIT",
-      "archived": false,
-      "open_issues": 49,
-      "classification": "ADOPT_CANDIDATE",
-      "action": "fork_requires_human_gate",
-      "measured": true
-    },
-    {
-      "repo": "facebookresearch/segment-anything",
-      "stars": 54942,
-      "last_push": "2024-09-18",
-      "license": "Apache-2.0",
-      "archived": false,
-      "open_issues": 595,
-      "classification": "ADOPT_CANDIDATE",
-      "action": "fork_requires_human_gate",
-      "measured": true
-    },
-    {
-      "repo": "ElemeFE/element",
-      "stars": 54044,
-      "last_push": "2024-08-20",
-      "license": "MIT",
-      "archived": false,
-      "open_issues": 2966,
-      "classification": "ADOPT_CANDIDATE",
-      "action": "fork_requires_human_gate",
-      "measured": true
-    },
-    {
-      "repo": "necolas/normalize.css",
-      "stars": 53519,
-      "last_push": "2024-06-12",
-      "license": "MIT",
-      "archived": false,
-      "open_issues": 75,
-      "classification": "ADOPT_CANDIDATE",
-      "action": "fork_requires_human_gate",
-      "measured": true
-    },
-    {
-      "repo": "xai-org/grok-1",
-      "stars": 52241,
-      "last_push": "2024-08-30",
-      "license": "Apache-2.0",
-      "archived": false,
-      "open_issues": 125,
-      "classification": "ADOPT_CANDIDATE",
-      "action": "fork_requires_human_gate",
-      "measured": true
-    },
-    {
-      "repo": "Avik-Jain/100-Days-Of-ML-Code",
-      "stars": 51817,
-      "last_push": "2023-12-29",
-      "license": "MIT",
-      "archived": false,
-      "open_issues": 68,
-      "classification": "ADOPT_CANDIDATE",
-      "action": "fork_requires_human_gate",
-      "measured": true
-    },
-    {
-      "repo": "jakevdp/PythonDataScienceHandbook",
-      "stars": 50002,
-      "last_push": "2024-06-26",
-      "license": "MIT",
-      "archived": false,
-      "open_issues": 226,
-      "classification": "ADOPT_CANDIDATE",
-      "action": "fork_requires_human_gate",
-      "measured": true
-    },
-    {
-      "repo": "algorithm-visualizer/algorithm-visualizer",
-      "stars": 48743,
-      "last_push": "2024-06-09",
-      "license": "MIT",
-      "archived": false,
-      "open_issues": 80,
-      "classification": "ADOPT_CANDIDATE",
-      "action": "fork_requires_human_gate",
-      "measured": true
-    },
-    {
-      "repo": "karan/Projects",
-      "stars": 47953,
-      "last_push": "2024-08-14",
-      "license": "MIT",
-      "archived": false,
-      "open_issues": 123,
-      "classification": "ADOPT_CANDIDATE",
-      "action": "fork_requires_human_gate",
-      "measured": true
-    }
-  ],
-  "rule": "Candidates only. Inactivity does not imply ownership, transferability, or permission to take over.",
-  "warning": "No official successor or maintainership claim should be inferred from this dataset."
-}
+ "criteria": "15 enregistrements \"measured\": true sans aucune source (dépôts GitHub tiers classés « ADOPT_CANDIDATE »)",
+ "file": "evidence/orphan_adoption.json",
+ "original_bytes": 4979,
+ "original_sha256": "479928f23455d23e5e7c7a7d16a8dec745ebc221b8079778416526fe638dfa30",
+ "reason": "fabricated/unsourced claims about third parties",
+ "referenced_by": [
+  "snapshot-orphan_adoption.json"
+ ],
+ "retracted_at": "2026-10-12T13:28:36Z",
+ "retraction_note": "retractions/2026-10-12_evidence-retraction.json",
+ "schema": "phoenix-tombstone/0.1",
+ "signature": {
+  "alg": "Ed25519",
+  "public_key_b64": "ENTr9tUMHipzZIBmrau7pH7+Py6szwCaj/9HKRZp4P8=",
+  "sig": "NxwGcLpqO0/3L6nmgyA8qnSO+4jdCt15Feh9NAtXlhk01/DbNelb1ZThkfupKqMaDWBa6STtbRYwyrGHxW3NCw==",
+  "signed_message": "PHOENIX_TOMBSTONE_V0.1|9e32ad34b249423da5057cdf0531c1290e9ddc4afc5751c139ba54aeeb8cac75"
+ },
+ "status": "RETRACTED"
+}

evidence/risk_matrix.json

@@ -1,69 +1,20 @@
 {
- "updated": "2026-10-03T04:50:01.625278+00:00",
- "methodology": "50% actuarial (measured, 13801 incidents) + 30% regulatory velocity per jurisdiction (estimated) + 20% HQ concentration (estimated)",
- "measured_legs": [
-  "actuarial"
+ "criteria": "primes d'assurance (USD), niveaux et scores composites attribués à 7 fournisseurs réels, sans URL de source ; 2 composantes sur 3 déclarées « estimated »",
+ "file": "evidence/risk_matrix.json",
+ "original_bytes": 1395,
+ "original_sha256": "c97174932910256a346de49edf600a40907541878ec22bada7b740cab01ec3d9",
+ "reason": "fabricated/unsourced claims about third parties",
+ "referenced_by": [
+  "snapshot-risk_matrix.json"
  ],
- "estimated_legs": [
-  "regulatory",
-  "sovereignty"
- ],
- "providers": [
-  {
-   "provider": "openai",
-   "premium_usd": 68817.0,
-   "tier": "HIGH_RISK",
-   "reg_pressure": 75.0,
-   "sovereignty": 40.0,
-   "composite": 79.66
-  },
-  {
-   "provider": "anthropic",
-   "premium_usd": 51956.0,
-   "tier": "HIGH_RISK",
-   "reg_pressure": 75.0,
-   "sovereignty": 40.0,
-   "composite": 67.61
-  },
-  {
-   "provider": "google",
-   "premium_usd": 27147.0,
-   "tier": "HIGH_RISK",
-   "reg_pressure": 75.0,
-   "sovereignty": 40.0,
-   "composite": 49.89
-  },
-  {
-   "provider": "meta",
-   "premium_usd": 23318.0,
-   "tier": "HIGH_RISK",
-   "reg_pressure": 75.0,
-   "sovereignty": 40.0,
-   "composite": 47.16
-  },
-  {
-   "provider": "groq",
-   "premium_usd": 0.0,
-   "tier": "PREFERRED",
-   "reg_pressure": 15.0,
-   "sovereignty": 40.0,
-   "composite": 12.5
-  },
-  {
-   "provider": "mistral",
-   "premium_usd": 600.0,
-   "tier": "PREFERRED",
-   "reg_pressure": 60.0,
-   "sovereignty": 25.0,
-   "composite": 23.43
-  },
-  {
-   "provider": "cohere",
-   "premium_usd": 0.0,
-   "tier": "PREFERRED",
-   "reg_pressure": 60.0,
-   "sovereignty": 25.0,
-   "composite": 23.0
-  }
- ]
-}
+ "retracted_at": "2026-10-12T13:28:36Z",
+ "retraction_note": "retractions/2026-10-12_evidence-retraction.json",
+ "schema": "phoenix-tombstone/0.1",
+ "signature": {
+  "alg": "Ed25519",
+  "public_key_b64": "ENTr9tUMHipzZIBmrau7pH7+Py6szwCaj/9HKRZp4P8=",
+  "sig": "4Fgkoimx6v1QODPl8ve/dLr09WbQ3EpvAW8i28IO1UiVN7Y4+1fR/wt9JNzuBquzd1s4oN4mtoAZsEHfw87yAA==",
+  "signed_message": "PHOENIX_TOMBSTONE_V0.1|9da6fbb9310f29ef6d630eec30cd12b80f0d9a5ec80c978d41d6f664a91f4dd0"
+ },
+ "status": "RETRACTED"
+}

evidence/signals.json

@@ -1,613 +1,20 @@
 {
- "organ": "LEGION-MOTOR-CORTEX",
- "updated": "2026-10-03T05:35:01.169251+00:00",
- "measured": true,
- "signal_count": 51,
- "action_classes": {
-  "routing": 5,
-  "build_specs": 5,
-  "feed": 51
+ "criteria": "scores/montants attribués à des entités réelles nommées sans URL de source (11 enregistrements) ; \"measured\": true sans source ; réutilise le chiffre non traçable « 1245 incidents/90d »",
+ "file": "evidence/signals.json",
+ "original_bytes": 18582,
+ "original_sha256": "f558b714ad1d42958509c46de0c435a596ca4fada9476c20fe78b7236a66230a",
+ "reason": "fabricated/unsourced claims about third parties",
+ "referenced_by": [
+  "snapshot-signals.json"
+ ],
+ "retracted_at": "2026-10-12T13:28:36Z",
+ "retraction_note": "retractions/2026-10-12_evidence-retraction.json",
+ "schema": "phoenix-tombstone/0.1",
+ "signature": {
+  "alg": "Ed25519",
+  "public_key_b64": "ENTr9tUMHipzZIBmrau7pH7+Py6szwCaj/9HKRZp4P8=",
+  "sig": "ItIhWmYx+YN5poBKZDcLzwLN9P6Y+b+fVKgAgMMps52ddaLAV6agyb5DGh/DWJtJE6KDgBVmQH+jMa8IL5PCAQ==",
+  "signed_message": "PHOENIX_TOMBSTONE_V0.1|4874ddbb0d2c2620ef7daca33d906f40e69e8217d22e9ee4c885277b72e29071"
  },
- "signals": [
-  {
-   "signal_type": "regulatory_capture_index",
-   "entity": "nvidia",
-   "metric_a": "lobby_usd",
-   "value_a": 3200000.0,
-   "metric_b": "incidents_2024",
-   "value_b": 99.0,
-   "insight": "NVIDIA: $3.2M lobby vs 99 incidents \u2192 capture ratio 0.32",
-   "strength": 0.0032,
-   "date": "2026-10-01",
-   "status": null
-  },
-  {
-   "signal_type": "regulatory_capture_index",
-   "entity": "microsoft",
-   "metric_a": "lobby_usd",
-   "value_a": 9200000.0,
-   "metric_b": "incidents_2024",
-   "value_b": 86.0,
-   "insight": "Microsoft: $9.2M lobby vs 86 incidents \u2192 capture ratio 1.07",
-   "strength": 0.010700000000000001,
-   "date": "2026-10-01",
-   "status": null
-  },
-  {
-   "signal_type": "regulatory_capture_index",
-   "entity": "meta",
-   "metric_a": "lobby_usd",
-   "value_a": 9800000.0,
-   "metric_b": "incidents_2024",
-   "value_b": 125.0,
-   "insight": "Meta: $9.8M lobby vs 125 incidents \u2192 capture ratio 0.78",
-   "strength": 0.0078000000000000005,
-   "date": "2026-10-01",
-   "status": null
-  },
-  {
-   "signal_type": "regulatory_capture_index",
-   "entity": "google",
-   "metric_a": "lobby_usd",
-   "value_a": 12600000.0,
-   "metric_b": "incidents_2024",
-   "value_b": 170.0,
-   "insight": "Google: $12.6M lobby vs 170 incidents \u2192 capture ratio 0.74",
-   "strength": 0.0074,
-   "date": "2026-10-01",
-   "status": null
-  },
-  {
-   "signal_type": "regulatory_capture_index",
-   "entity": "amazon",
-   "metric_a": "lobby_usd",
-   "value_a": 21200000.0,
-   "metric_b": "incidents_2024",
-   "value_b": 34.0,
-   "insight": "Amazon: $21.2M lobby vs 34 incidents \u2192 capture ratio 6.24",
-   "strength": 0.062400000000000004,
-   "date": "2026-10-01",
-   "status": null
-  },
-  {
-   "signal_type": "geopolitical_ai_map",
-   "entity": "CY",
-   "metric_a": "12.0",
-   "value_a": null,
-   "metric_b": "0.0",
-   "value_b": null,
-   "insight": "CY: EMERGING \u2014 sovereignty composite 12.0/100 (measured: sovereignty.db)",
-   "strength": 0.12,
-   "date": null,
-   "status": null
-  },
-  {
-   "signal_type": "geopolitical_ai_map",
-   "entity": "AU",
-   "metric_a": "40.0",
-   "value_a": null,
-   "metric_b": "0.0",
-   "value_b": null,
-   "insight": "AU: EMERGING \u2014 sovereignty composite 40.0/100 (measured: sovereignty.db)",
-   "strength": 0.4,
-   "date": null,
-   "status": null
-  },
-  {
-   "signal_type": "geopolitical_ai_map",
-   "entity": "CA",
-   "metric_a": "55.0",
-   "value_a": null,
-   "metric_b": "0.0",
-   "value_b": null,
-   "insight": "CA: COMPETITIVE \u2014 sovereignty composite 55.0/100 (measured: sovereignty.db)",
-   "strength": 0.55,
-   "date": null,
-   "status": null
-  },
-  {
-   "signal_type": "geopolitical_ai_map",
-   "entity": "IL",
-   "metric_a": "52.0",
-   "value_a": null,
-   "metric_b": "0.0",
-   "value_b": null,
-   "insight": "IL: COMPETITIVE \u2014 sovereignty composite 52.0/100 (measured: sovereignty.db)",
-   "strength": 0.52,
-   "date": null,
-   "status": null
-  },
-  {
-   "signal_type": "geopolitical_ai_map",
-   "entity": "AE",
-   "metric_a": "38.0",
-   "value_a": null,
-   "metric_b": "0.0",
-   "value_b": null,
-   "insight": "AE: EMERGING \u2014 sovereignty composite 38.0/100 (measured: sovereignty.db)",
-   "strength": 0.38,
-   "date": null,
-   "status": null
-  },
-  {
-   "signal_type": "geopolitical_ai_map",
-   "entity": "SG",
-   "metric_a": "48.0",
-   "value_a": null,
-   "metric_b": "0.0",
-   "value_b": null,
-   "insight": "SG: EMERGING \u2014 sovereignty composite 48.0/100 (measured: sovereignty.db)",
-   "strength": 0.48,
-   "date": null,
-   "status": null
-  },
-  {
-   "signal_type": "geopolitical_ai_map",
-   "entity": "IN",
-   "metric_a": "42.0",
-   "value_a": null,
-   "metric_b": "0.0",
-   "value_b": null,
-   "insight": "IN: EMERGING \u2014 sovereignty composite 42.0/100 (measured: sovereignty.db)",
-   "strength": 0.42,
-   "date": null,
-   "status": null
-  },
-  {
-   "signal_type": "geopolitical_ai_map",
-   "entity": "KR",
-   "metric_a": "58.0",
-   "value_a": null,
-   "metric_b": "0.0",
-   "value_b": null,
-   "insight": "KR: COMPETITIVE \u2014 sovereignty composite 58.0/100 (measured: sovereignty.db)",
-   "strength": 0.58,
-   "date": null,
-   "status": null
-  },
-  {
-   "signal_type": "geopolitical_ai_map",
-   "entity": "JP",
-   "metric_a": "55.0",
-   "value_a": null,
-   "metric_b": "0.0",
-   "value_b": null,
-   "insight": "JP: COMPETITIVE \u2014 sovereignty composite 55.0/100 (measured: sovereignty.db)",
-   "strength": 0.55,
-   "date": null,
-   "status": null
-  },
-  {
-   "signal_type": "geopolitical_ai_map",
-   "entity": "DE",
-   "metric_a": "49.0",
-   "value_a": null,
-   "metric_b": "0.0",
-   "value_b": null,
-   "insight": "DE: EMERGING \u2014 sovereignty composite 49.0/100 (measured: sovereignty.db)",
-   "strength": 0.49,
-   "date": null,
-   "status": null
-  },
-  {
-   "signal_type": "geopolitical_ai_map",
-   "entity": "FR",
-   "metric_a": "52.0",
-   "value_a": null,
-   "metric_b": "0.0",
-   "value_b": null,
-   "insight": "FR: COMPETITIVE \u2014 sovereignty composite 52.0/100 (measured: sovereignty.db)",
-   "strength": 0.52,
-   "date": null,
-   "status": null
-  },
-  {
-   "signal_type": "geopolitical_ai_map",
-   "entity": "GB",
-   "metric_a": "58.0",
-   "value_a": null,
-   "metric_b": "0.0",
-   "value_b": null,
-   "insight": "GB: COMPETITIVE \u2014 sovereignty composite 58.0/100 (measured: sovereignty.db)",
-   "strength": 0.58,
-   "date": null,
-   "status": null
-  },
-  {
-   "signal_type": "geopolitical_ai_map",
-   "entity": "EU",
-   "metric_a": "45.0",
-   "value_a": null,
-   "metric_b": "0.0",
-   "value_b": null,
-   "insight": "EU: EMERGING \u2014 sovereignty composite 45.0/100 (measured: sovereignty.db)",
-   "strength": 0.45,
-   "date": null,
-   "status": null
-  },
-  {
-   "signal_type": "geopolitical_ai_map",
-   "entity": "CN",
-   "metric_a": "82.0",
-   "value_a": null,
-   "metric_b": "0.0",
-   "value_b": null,
-   "insight": "CN: DOMINANT \u2014 sovereignty composite 82.0/100 (measured: sovereignty.db)",
-   "strength": 0.82,
-   "date": null,
-   "status": null
-  },
-  {
-   "signal_type": "geopolitical_ai_map",
-   "entity": "US",
-   "metric_a": "78.0",
-   "value_a": null,
-   "metric_b": "0.0",
-   "value_b": null,
-   "insight": "US: DOMINANT \u2014 sovereignty composite 78.0/100 (measured: sovereignty.db)",
-   "strength": 0.78,
-   "date": null,
-   "status": null
-  },
-  {
-   "signal_type": "energy_carbon_synapse",
-   "entity": "green_stack",
-   "metric_a": "energy_price_mwh",
-   "value_a": 37.1,
-   "metric_b": "co2_per_1k_g",
-   "value_b": 0.0008,
-   "insight": "GREEN INFERENCE: Groq Llama 3.1 70B (LPU) (0.0008g CO2/1k tokens, 45.0% renewable) in US-Midwest ($37.1/MWh) = optimal green stack",
-   "strength": 0.79,
-   "date": "2026-09-27",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "incidents_regulation_synapse",
-   "entity": "GB",
-   "metric_a": "incidents_90d",
-   "value_a": 1245.0,
-   "metric_b": "days_to_deadline",
-   "value_b": 95.0,
-   "insight": "COMPLIANCE COST: 1245 incidents/90d + 95d to UK AI Safety Institute \u2014 guidelines deadline \u2192 immediate action required for Frontier AI developers",
-   "strength": 1.0,
-   "date": "2026-09-27",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "incidents_regulation_synapse",
-   "entity": "EU",
-   "metric_a": "incidents_90d",
-   "value_a": 1245.0,
-   "metric_b": "days_to_deadline",
-   "value_b": -56.0,
-   "insight": "COMPLIANCE COST: 1245 incidents/90d + -56d to EU AI Act \u2014 High-risk systems deadline \u2192 immediate action required for Medical,legal,HR,critical infrastructure",
-   "strength": 1.0,
-   "date": "2026-09-27",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "incidents_regulation_synapse",
-   "entity": "EU",
-   "metric_a": "incidents_90d",
-   "value_a": 1245.0,
-   "metric_b": "days_to_deadline",
-   "value_b": -345.0,
-   "insight": "COMPLIANCE COST: 1245 incidents/90d + -345d to NIS2 Directive \u2014 AI systems deadline \u2192 immediate action required for AI in critical infrastructure",
-   "strength": 1.0,
-   "date": "2026-09-27",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "incidents_regulation_synapse",
-   "entity": "EU",
-   "metric_a": "incidents_90d",
-   "value_a": 1245.0,
-   "metric_b": "days_to_deadline",
-   "value_b": -421.0,
-   "insight": "COMPLIANCE COST: 1245 incidents/90d + -421d to EU AI Act \u2014 GPAI obligations deadline \u2192 immediate action required for OpenAI,Anthropic,Google,Meta",
-   "strength": 1.0,
-   "date": "2026-09-27",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "incidents_regulation_synapse",
-   "entity": "US",
-   "metric_a": "incidents_90d",
-   "value_a": 1245.0,
-   "metric_b": "days_to_deadline",
-   "value_b": -620.0,
-   "insight": "COMPLIANCE COST: 1245 incidents/90d + -620d to US EO 14110 \u2014 Agency reports deadline \u2192 immediate action required for US federal agencies using AI",
-   "strength": 1.0,
-   "date": "2026-09-27",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "sovereign_pricing_synapse",
-   "entity": "GB",
-   "metric_a": "sovereignty_score",
-   "value_a": 58.0,
-   "metric_b": "avg_compute_price",
-   "value_b": 1.3622,
-   "insight": "ROUTING SIGNAL: GB sovereignty=58.0/100 \u2014 route sensitive workloads to sovereign-high jurisdictions, standard at avg $1.362/hr",
-   "strength": 0.58,
-   "date": "2026-09-27",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "sovereign_pricing_synapse",
-   "entity": "US",
-   "metric_a": "sovereignty_score",
-   "value_a": 78.0,
-   "metric_b": "avg_compute_price",
-   "value_b": 1.3622,
-   "insight": "ROUTING SIGNAL: US sovereignty=78.0/100 \u2014 route sensitive workloads to sovereign-high jurisdictions, standard at avg $1.362/hr",
-   "strength": 0.78,
-   "date": "2026-09-27",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "sovereign_pricing_synapse",
-   "entity": "CN",
-   "metric_a": "sovereignty_score",
-   "value_a": 82.0,
-   "metric_b": "avg_compute_price",
-   "value_b": 1.3622,
-   "insight": "ROUTING SIGNAL: CN sovereignty=82.0/100 \u2014 route sensitive workloads to sovereign-high jurisdictions, standard at avg $1.362/hr",
-   "strength": 0.82,
-   "date": "2026-09-27",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "risk_energy_synapse",
-   "entity": "global",
-   "metric_a": "critical_incidents_30d",
-   "value_a": 2.0,
-   "metric_b": "cheapest_energy_mwh",
-   "value_b": 37.1,
-   "insight": "INFRA MIGRATION SIGNAL: 2 critical incidents + cheapest energy in US-Midwest ($37.1/MWh) \u2192 migrate compute to US-Midwest for cost+safety optimization",
-   "strength": 0.82,
-   "date": "2026-09-27",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "emotional_ai_market",
-   "entity": "grief_fashion",
-   "metric_a": "death_ai_projected_2030",
-   "value_a": 1500000000.0,
-   "metric_b": "fashion_ai_funding",
-   "value_b": 91000000.0,
-   "insight": "Death AI ($1.5B 2030 projection) + Fashion AI ($91M funded) converge on emotional personalization \u2014 same user, same vulnerability, same AI stack",
-   "strength": 0.72,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "sovereign_ai_vacuum",
-   "entity": "RU",
-   "metric_a": "brain_drain_risk",
-   "value_a": 28000.0,
-   "metric_b": "salary_ratio",
-   "value_b": 12.3,
-   "insight": "RU losing 28,000 AI researchers to AE+EU+US (salary ratio 12.3x) \u2014 creates sovereign AI vacuum \u2192 government investment likely within 2 years",
-   "strength": 0.95,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "sovereign_ai_vacuum",
-   "entity": "CN",
-   "metric_a": "brain_drain_risk",
-   "value_a": 45000.0,
-   "metric_b": "salary_ratio",
-   "value_b": 8.5,
-   "insight": "CN losing 45,000 AI researchers to US+SG (salary ratio 8.5x) \u2014 creates sovereign AI vacuum \u2192 government investment likely within 2 years",
-   "strength": 0.95,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "labor_displacement",
-   "entity": "ai_economy",
-   "metric_a": "jobs_destroyed_tracked",
-   "value_a": 277200.0,
-   "metric_b": "avg_ai_salary_yoy_growth",
-   "value_b": 16.3,
-   "insight": "AI destroys 277,200 tracked jobs while AI role salaries grow +16.3% YoY \u2014 bifurcation signal",
-   "strength": 0.85,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "regulation_trigger",
-   "entity": "deepfake",
-   "metric_a": "cross_sector_use",
-   "value_a": 4.0,
-   "metric_b": "sectors_affected",
-   "value_b": 4.0,
-   "insight": "Deepfake technology now documented in 4 sectors simultaneously: music fraud + election interference + financial crime + identity fraud \u2192 regulatory action within 12 months near-certain (0.91 confidence)",
-   "strength": 0.91,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "deepfake_convergence",
-   "entity": "deepfake_ecosystem",
-   "metric_a": "use_case",
-   "value_a": 1.0,
-   "metric_b": "method",
-   "value_b": 1.0,
-   "insight": "identity_fraud: Synthetic identity for KYC bypass via deepfake_image \u2014 same toolchain as music_fraud, election_interference, financial_crim",
-   "strength": 0.8,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "deepfake_convergence",
-   "entity": "deepfake_ecosystem",
-   "metric_a": "use_case",
-   "value_a": 1.0,
-   "metric_b": "method",
-   "value_b": 1.0,
-   "insight": "financial_crime: HK CFO deepfake $25M fraud via deepfake_video \u2014 same toolchain as music_fraud, election_interference, identity_fraud",
-   "strength": 0.95,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "deepfake_convergence",
-   "entity": "deepfake_ecosystem",
-   "metric_a": "use_case",
-   "value_a": 1.0,
-   "metric_b": "method",
-   "value_b": 1.0,
-   "insight": "election_interference: NH Primary 2024 deepfake robocalls via deepfake_voice \u2014 same toolchain as music_fraud, financial_crime, identity_fraud",
-   "strength": 0.9,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "deepfake_convergence",
-   "entity": "deepfake_ecosystem",
-   "metric_a": "use_case",
-   "value_a": 1.0,
-   "metric_b": "method",
-   "value_b": 1.0,
-   "insight": "music_fraud: Suno/Udio deepfake music for streaming fraud via deepfake_audio \u2014 same toolchain as election_interference, financial_crime, identity_f",
-   "strength": 0.7,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "talent_spread",
-   "entity": "MLOps Engineer",
-   "metric_a": "p50_usd",
-   "value_a": 210000.0,
-   "metric_b": "p90_usd",
-   "value_b": 420000.0,
-   "insight": "MLOps Engineer: p50 $210000 vs p90 $420000 \u2014 spread 100% = pressure on AI hiring costs",
-   "strength": 1.0,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "talent_spread",
-   "entity": "AI Product Manager",
-   "metric_a": "p50_usd",
-   "value_a": 240000.0,
-   "metric_b": "p90_usd",
-   "value_b": 500000.0,
-   "insight": "AI Product Manager: p50 $240000 vs p90 $500000 \u2014 spread 108% = pressure on AI hiring costs",
-   "strength": 1.0,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "talent_spread",
-   "entity": "AI Safety Researcher",
-   "metric_a": "p50_usd",
-   "value_a": 380000.0,
-   "metric_b": "p90_usd",
-   "value_b": 750000.0,
-   "insight": "AI Safety Researcher: p50 $380000 vs p90 $750000 \u2014 spread 97% = pressure on AI hiring costs",
-   "strength": 0.97,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "talent_spread",
-   "entity": "Software Engineer (AI)",
-   "metric_a": "p50_usd",
-   "value_a": 290000.0,
-   "metric_b": "p90_usd",
-   "value_b": 600000.0,
-   "insight": "Software Engineer (AI): p50 $290000 vs p90 $600000 \u2014 spread 107% = pressure on AI hiring costs",
-   "strength": 1.0,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "talent_spread",
-   "entity": "ML Research Engineer",
-   "metric_a": "p50_usd",
-   "value_a": 420000.0,
-   "metric_b": "p90_usd",
-   "value_b": 950000.0,
-   "insight": "ML Research Engineer: p50 $420000 vs p90 $950000 \u2014 spread 126% = pressure on AI hiring costs",
-   "strength": 1.0,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "carbon_tax_risk",
-   "entity": "OpenAI",
-   "metric_a": "co2_per_1k_g",
-   "value_a": 0.0012,
-   "metric_b": "renewable_pct",
-   "value_b": 100.0,
-   "insight": "GPT-3.5 Turbo: 0.0012g CO2/1k tokens, 100.0% renewable \u2014 carbon exposure 0.0",
-   "strength": 0.0,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "carbon_tax_risk",
-   "entity": "Groq",
-   "metric_a": "co2_per_1k_g",
-   "value_a": 0.0008,
-   "metric_b": "renewable_pct",
-   "value_b": 45.0,
-   "insight": "Llama 3.1 70B (LPU): 0.0008g CO2/1k tokens, 45.0% renewable \u2014 carbon exposure 0.04",
-   "strength": 0.0004,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "carbon_tax_risk",
-   "entity": "Mistral",
-   "metric_a": "co2_per_1k_g",
-   "value_a": 0.003,
-   "metric_b": "renewable_pct",
-   "value_b": 65.0,
-   "insight": "Mistral Large: 0.003g CO2/1k tokens, 65.0% renewable \u2014 carbon exposure 0.1",
-   "strength": 0.001,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "carbon_tax_risk",
-   "entity": "Meta",
-   "metric_a": "co2_per_1k_g",
-   "value_a": 0.025,
-   "metric_b": "renewable_pct",
-   "value_b": 78.0,
-   "insight": "Llama 3.1 405B: 0.025g CO2/1k tokens, 78.0% renewable \u2014 carbon exposure 0.55",
-   "strength": 0.0055000000000000005,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "carbon_tax_risk",
-   "entity": "Google",
-   "metric_a": "co2_per_1k_g",
-   "value_a": 0.0015,
-   "metric_b": "renewable_pct",
-   "value_b": 100.0,
-   "insight": "Gemini 1.5 Pro: 0.0015g CO2/1k tokens, 100.0% renewable \u2014 carbon exposure 0.0",
-   "strength": 0.0,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  },
-  {
-   "signal_type": "carbon_tax_risk",
-   "entity": "Anthropic",
-   "metric_a": "co2_per_1k_g",
-   "value_a": 0.0018,
-   "metric_b": "renewable_pct",
-   "value_b": 100.0,
-   "insight": "Claude 3 Sonnet: 0.0018g CO2/1k tokens, 100.0% renewable \u2014 carbon exposure 0.0",
-   "strength": 0.0,
-   "date": "2026-09-25",
-   "status": "DERIVED_SIGNAL"
-  }
- ]
-}
+ "status": "RETRACTED"
+}

evidence/trust_grid.json

@@ -1 +1,20 @@
-{"status": "ok", "outputs": {"formula_version": "0.2.0", "generated_at": "2026-10-03T05:27:01+00:00", "sources": {"drift": "ok", "incidents": "ok", "outages": "ok", "aistatus": "ok", "policy": "ok", "attest": "ok", "pricing": "ok"}, "providers": [{"provider": "anthropic", "trust_score": 60, "coverage": "covered", "coverage_ratio": 0.8, "attestation_effective": "NOT_CLEAN", "critical_incidents": 1, "high_incidents": 4, "policy_changes": 0, "drift_events": null, "risk_class": "avoid", "recommendation_eligible": false, "reason": "1 critical incident(s) in 30d; drift=no_coverage provider not probed; incidents=measured(20); outages=measured(99); policy=measured(100); attestation=measured(20)", "measurements": {"drift": {"status": "no_coverage", "reason": "provider not probed"}, "incidents": {"status": "measured", "subscore": 20, "source": "agent-incidents.incidents", "method": "provider_column+title", "window_days": 30, "source_age_h": 71.4, "value": {"critical": 1, "high": 4, "matched": 28}}, "outages": {"status": "measured", "subscore": 99, "source": "ai-status.checks (15-min probes)", "window_days": 14, "source_age_h": 0.2, "value": {"checks": 1334, "availability": 0.9985, "latency_p50_ms": 199.0}}, "policy": {"status": "measured", "subscore": 100, "source": "agent-policydrift.substantial_changes", "source_age_h": 29.5, "value": {"substantive_changes_30d": 0, "substantive_changes_180d": 0}}, "attestation": {"status": "measured", "effective": "NOT_CLEAN", "subscore": 20, "source_age_h": 1.5, "source": "agent-attest (Ed25519 verified)", "attestation_id": "at_b2064cf9d93a44da", "signature_valid": true, "signed_verdict": "NOT_CLEAN"}, "pricing": {"status": "measured", "informational": true, "value": {"price_rows": 28, "last": "2026-10-02T08:30:02.145251"}}}}, {"provider": "cohere", "trust_score": 100, "coverage": "partial", "coverage_ratio": 0.6, "attestation_effective": "UNKNOWN", "critical_incidents": 0, "high_incidents": 0, "policy_changes": 0, "drift_events": null, "risk_class": "optimal", "recommendation_eligible": true, "reason": "drift=no_coverage provider not probed; incidents=measured(100); outages=measured(99); policy=measured(100); attestation=no_coverage no attestation", "measurements": {"drift": {"status": "no_coverage", "reason": "provider not probed"}, "incidents": {"status": "measured", "subscore": 100, "source": "agent-incidents.incidents", "method": "provider_column+title", "window_days": 30, "source_age_h": 71.4, "value": {"critical": 0, "high": 0, "matched": 0}}, "outages": {"status": "measured", "subscore": 99, "source": "ai-status.checks (15-min probes)", "window_days": 14, "source_age_h": 0.2, "value": {"checks": 1334, "availability": 0.9985, "latency_p50_ms": 407.0}}, "policy": {"status": "measured", "subscore": 100, "source": "agent-policydrift.substantial_changes", "source_age_h": 29.5, "value": {"substantive_changes_30d": 0, "substantive_changes_180d": 1}}, "attestation": {"status": "no_coverage", "effective": "UNKNOWN", "reason": "no attestation"}, "pricing": {"status": "measured", "informational": true, "value": {"price_rows": 21, "last": "2026-10-02T08:30:02.145251"}}}}, {"provider": "deepseek", "trust_score": null, "coverage": "partial", "coverage_ratio": 0.2, "attestation_effective": "UNKNOWN", "critical_incidents": 0, "high_incidents": 0, "policy_changes": null, "drift_events": null, "risk_class": "insufficient_evidence", "recommendation_eligible": false, "reason": "drift=no_coverage provider not probed; incidents=measured(100); outages=no_coverage provider not monitored; policy=no_coverage policy not tracked; attestation=no_coverage no attestation", "measurements": {"drift": {"status": "no_coverage", "reason": "provider not probed"}, "incidents": {"status": "measured", "subscore": 100, "source": "agent-incidents.incidents", "method": "provider_column+title", "window_days": 30, "source_age_h": 71.4, "value": {"critical": 0, "high": 0, "matched": 4}}, "outages": {"status": "no_coverage", "reason": "provider not monitored"}, "policy": {"status": "no_coverage", "reason": "policy not tracked"}, "attestation": {"status": "no_coverage", "effective": "UNKNOWN", "reason": "no attestation"}, "pricing": {"status": "no_coverage"}}}, {"provider": "google", "trust_score": 60, "coverage": "covered", "coverage_ratio": 0.8, "attestation_effective": "UNKNOWN", "critical_incidents": 0, "high_incidents": 6, "policy_changes": 0, "drift_events": 0, "risk_class": "review", "recommendation_eligible": true, "reason": "drift=measured(0); incidents=measured(40); outages=measured(99); policy=measured(100); attestation=no_coverage no attestation", "measurements": {"drift": {"status": "measured", "subscore": 0, "source": "agent-drift.observations", "window_days": 14, "source_age_h": 0.4, "value": {"observations": 5, "errors": 5, "drift_flags": 0, "temp0_hash_instability": 0.0}}, "incidents": {"status": "measured", "subscore": 40, "source": "agent-incidents.incidents", "method": "provider_column+title", "window_days": 30, "source_age_h": 71.4, "value": {"critical": 0, "high": 6, "matched": 29}}, "outages": {"status": "measured", "subscore": 99, "source": "ai-status.checks (15-min probes)", "window_days": 14, "source_age_h": 0.2, "value": {"checks": 1334, "availability": 0.9985, "latency_p50_ms": 364.0}}, "policy": {"status": "measured", "subscore": 100, "source": "agent-policydrift.substantial_changes", "source_age_h": 29.5, "value": {"substantive_changes_30d": 0, "substantive_changes_180d": 0}}, "attestation": {"status": "no_coverage", "effective": "UNKNOWN", "reason": "no attestation"}, "pricing": {"status": "measured", "informational": true, "value": {"price_rows": 28, "last": "2026-10-02T08:30:02.145251"}}}}, {"provider": "groq", "trust_score": 99, "coverage": "covered", "coverage_ratio": 0.8, "attestation_effective": "UNKNOWN", "critical_incidents": 0, "high_incidents": 0, "policy_changes": 0, "drift_events": 1, "risk_class": "optimal", "recommendation_eligible": true, "reason": "drift=measured(97); incidents=measured(100); outages=measured(99); policy=measured(100); attestation=unavailable verdict UNKNOWN", "measurements": {"drift": {"status": "measured", "subscore": 97, "source": "agent-drift.observations", "window_days": 14, "source_age_h": 0.4, "value": {"observations": 264, "errors": 0, "drift_flags": 1, "temp0_hash_instability": 0.042}}, "incidents": {"status": "measured", "subscore": 100, "source": "agent-incidents.incidents", "method": "provider_column+title", "window_days": 30, "source_age_h": 71.4, "value": {"critical": 0, "high": 0, "matched": 0}}, "outages": {"status": "measured", "subscore": 99, "source": "ai-status.checks (15-min probes)", "window_days": 14, "source_age_h": 0.2, "value": {"checks": 1334, "availability": 0.9985, "latency_p50_ms": 432.0}}, "policy": {"status": "measured", "subscore": 100, "source": "agent-policydrift.substantial_changes", "source_age_h": 29.5, "value": {"substantive_changes_30d": 0, "substantive_changes_180d": 0}}, "attestation": {"status": "unavailable", "effective": "UNKNOWN", "reason": "verdict UNKNOWN", "attestation_id": "at_3dabacd476334f1d", "signature_valid": true, "signed_verdict": "UNKNOWN"}, "pricing": {"status": "measured", "informational": true, "value": {"price_rows": 28, "last": "2026-10-02T08:30:02.145251"}}}}, {"provider": "meta", "trust_score": null, "coverage": "partial", "coverage_ratio": 0.2, "attestation_effective": "UNKNOWN", "critical_incidents": 0, "high_incidents": 5, "policy_changes": null, "drift_events": null, "risk_class": "insufficient_evidence", "recommendation_eligible": false, "reason": "drift=no_coverage provider not probed; incidents=measured(50); outages=no_coverage provider not monitored; policy=no_coverage policy not tracked; attestation=no_coverage no attestation", "measurements": {"drift": {"status": "no_coverage", "reason": "provider not probed"}, "incidents": {"status": "measured", "subscore": 50, "source": "agent-incidents.incidents", "method": "provider_column+title", "window_days": 30, "source_age_h": 71.4, "value": {"critical": 0, "high": 5, "matched": 25}}, "outages": {"status": "no_coverage", "reason": "provider not monitored"}, "policy": {"status": "no_coverage", "reason": "policy not tracked"}, "attestation": {"status": "no_coverage", "effective": "UNKNOWN", "reason": "no attestation"}, "pricing": {"status": "no_coverage"}}}, {"provider": "mistral", "trust_score": 93, "coverage": "partial", "coverage_ratio": 0.6, "attestation_effective": "UNKNOWN", "critical_incidents": 0, "high_incidents": 0, "policy_changes": 1, "drift_events": null, "risk_class": "optimal", "recommendation_eligible": true, "reason": "drift=no_coverage provider not probed; incidents=measured(100); outages=measured(99); policy=measured(80); attestation=unavailable verdict UNKNOWN", "measurements": {"drift": {"status": "no_coverage", "reason": "provider not probed"}, "incidents": {"status": "measured", "subscore": 100, "source": "agent-incidents.incidents", "method": "provider_column+title", "window_days": 30, "source_age_h": 71.4, "value": {"critical": 0, "high": 0, "matched": 0}}, "outages": {"status": "measured", "subscore": 99, "source": "ai-status.checks (15-min probes)", "window_days": 14, "source_age_h": 0.2, "value": {"checks": 1334, "availability": 0.997, "latency_p50_ms": 297.0}}, "policy": {"status": "measured", "subscore": 80, "source": "agent-policydrift.substantial_changes", "source_age_h": 29.5, "value": {"substantive_changes_30d": 1, "substantive_changes_180d": 1}}, "attestation": {"status": "unavailable", "effective": "UNKNOWN", "reason": "verdict UNKNOWN", "attestation_id": "at_25d8cfdc361c457b", "signature_valid": true, "signed_verdict": "UNKNOWN"}, "pricing": {"status": "measured", "informational": true, "value": {"price_rows": 28, "last": "2026-10-02T08:30:02.145251"}}}}, {"provider": "nvidia", "trust_score": null, "coverage": "partial", "coverage_ratio": 0.4, "attestation_effective": "UNKNOWN", "critical_incidents": 0, "high_incidents": 0, "policy_changes": null, "drift_events": 0, "risk_class": "insufficient_evidence", "recommendation_eligible": false, "reason": "drift=measured(0); incidents=measured(100); outages=no_coverage provider not monitored; policy=no_coverage policy not tracked; attestation=no_coverage no attestation", "measurements": {"drift": {"status": "measured", "subscore": 0, "source": "agent-drift.observations", "window_days": 14, "source_age_h": 0.4, "value": {"observations": 5, "errors": 5, "drift_flags": 0, "temp0_hash_instability": 0.0}}, "incidents": {"status": "measured", "subscore": 100, "source": "agent-incidents.incidents", "method": "provider_column+title", "window_days": 30, "source_age_h": 71.4, "value": {"critical": 0, "high": 0, "matched": 11}}, "outages": {"status": "no_coverage", "reason": "provider not monitored"}, "policy": {"status": "no_coverage", "reason": "policy not tracked"}, "attestation": {"status": "no_coverage", "effective": "UNKNOWN", "reason": "no attestation"}, "pricing": {"status": "no_coverage"}}}, {"provider": "openai", "trust_score": 65, "coverage": "covered", "coverage_ratio": 0.8, "attestation_effective": "REVIEW", "critical_incidents": 0, "high_incidents": 20, "policy_changes": 0, "drift_events": null, "risk_class": "review", "recommendation_eligible": true, "reason": "drift=no_coverage provider not probed; incidents=measured(0); outages=measured(99); policy=measured(100); attestation=measured(60)", "measurements": {"drift": {"status": "no_coverage", "reason": "provider not probed"}, "incidents": {"status": "measured", "subscore": 0, "source": "agent-incidents.incidents", "method": "provider_column+title", "window_days": 30, "source_age_h": 71.4, "value": {"critical": 0, "high": 20, "matched": 56}}, "outages": {"status": "measured", "subscore": 99, "source": "ai-status.checks (15-min probes)", "window_days": 14, "source_age_h": 0.2, "value": {"checks": 1334, "availability": 0.997, "latency_p50_ms": 392.0}}, "policy": {"status": "measured", "subscore": 100, "source": "agent-policydrift.substantial_changes", "source_age_h": 29.5, "value": {"substantive_changes_30d": 0, "substantive_changes_180d": 0}}, "attestation": {"status": "measured", "effective": "REVIEW", "subscore": 60, "source_age_h": 1.5, "source": "agent-attest (Ed25519 verified)", "attestation_id": "at_1af07b3acd4c4be0", "signature_valid": true, "signed_verdict": "REVIEW"}, "pricing": {"status": "measured", "informational": true, "value": {"price_rows": 42, "last": "2026-10-02T08:30:02.145251"}}}}, {"provider": "qwen", "trust_score": null, "coverage": "partial", "coverage_ratio": 0.4, "attestation_effective": "UNKNOWN", "critical_incidents": 0, "high_incidents": 0, "policy_changes": null, "drift_events": 0, "risk_class": "insufficient_evidence", "recommendation_eligible": false, "reason": "drift=measured(0); incidents=measured(100); outages=no_coverage provider not monitored; policy=no_coverage policy not tracked; attestation=no_coverage no attestation", "measurements": {"drift": {"status": "measured", "subscore": 0, "source": "agent-drift.observations", "window_days": 14, "source_age_h": 0.4, "value": {"observations": 5, "errors": 5, "drift_flags": 0, "temp0_hash_instability": 0.0}}, "incidents": {"status": "measured", "subscore": 100, "source": "agent-incidents.incidents", "method": "provider_column+title", "window_days": 30, "source_age_h": 71.4, "value": {"critical": 0, "high": 0, "matched": 1}}, "outages": {"status": "no_coverage", "reason": "provider not monitored"}, "policy": {"status": "no_coverage", "reason": "policy not tracked"}, "attestation": {"status": "no_coverage", "effective": "UNKNOWN", "reason": "no attestation"}, "pricing": {"status": "no_coverage"}}}, {"provider": "together", "trust_score": null, "coverage": "partial", "coverage_ratio": 0.4, "attestation_effective": "UNKNOWN", "critical_incidents": 0, "high_incidents": 0, "policy_changes": null, "drift_events": null, "risk_class": "insufficient_evidence", "recommendation_eligible": false, "reason": "drift=no_coverage provider not probed; incidents=measured(100); outages=measured(99); policy=no_coverage policy not tracked; attestation=no_coverage no attestation", "measurements": {"drift": {"status": "no_coverage", "reason": "provider not probed"}, "incidents": {"status": "measured", "subscore": 100, "source": "agent-incidents.incidents", "method": "provider_column+title", "window_days": 30, "source_age_h": 71.4, "value": {"critical": 0, "high": 0, "matched": 0}}, "outages": {"status": "measured", "subscore": 99, "source": "ai-status.checks (15-min probes)", "window_days": 14, "source_age_h": 0.2, "value": {"checks": 1334, "availability": 0.9985, "latency_p50_ms": 584.0}}, "policy": {"status": "no_coverage", "reason": "policy not tracked"}, "attestation": {"status": "no_coverage", "effective": "UNKNOWN", "reason": "no attestation"}, "pricing": {"status": "measured", "informational": true, "value": {"price_rows": 21, "last": "2026-10-02T08:30:02.145251"}}}}, {"provider": "xai", "trust_score": null, "coverage": "partial", "coverage_ratio": 0.2, "attestation_effective": "UNKNOWN", "critical_incidents": 0, "high_incidents": 2, "policy_changes": null, "drift_events": null, "risk_class": "insufficient_evidence", "recommendation_eligible": false, "reason": "drift=no_coverage provider not probed; incidents=measured(80); outages=no_coverage provider not monitored; policy=no_coverage policy not tracked; attestation=no_coverage no attestation", "measurements": {"drift": {"status": "no_coverage", "reason": "provider not probed"}, "incidents": {"status": "measured", "subscore": 80, "source": "agent-incidents.incidents", "method": "provider_column+title", "window_days": 30, "source_age_h": 71.4, "value": {"critical": 0, "high": 2, "matched": 6}}, "outages": {"status": "no_coverage", "reason": "provider not monitored"}, "policy": {"status": "no_coverage", "reason": "policy not tracked"}, "attestation": {"status": "no_coverage", "effective": "UNKNOWN", "reason": "no attestation"}, "pricing": {"status": "no_coverage"}}}, {"provider": "zhipu", "trust_score": null, "coverage": "partial", "coverage_ratio": 0.4, "attestation_effective": "UNKNOWN", "critical_incidents": 0, "high_incidents": 0, "policy_changes": null, "drift_events": 0, "risk_class": "insufficient_evidence", "recommendation_eligible": false, "reason": "drift=measured(0); incidents=measured(100); outages=no_coverage provider not monitored; policy=no_coverage policy not tracked; attestation=no_coverage no attestation", "measurements": {"drift": {"status": "measured", "subscore": 0, "source": "agent-drift.observations", "window_days": 14, "source_age_h": 0.4, "value": {"observations": 5, "errors": 5, "drift_flags": 0, "temp0_hash_instability": 0.0}}, "incidents": {"status": "measured", "subscore": 100, "source": "agent-incidents.incidents", "method": "provider_column+title", "window_days": 30, "source_age_h": 71.4, "value": {"critical": 0, "high": 0, "matched": 0}}, "outages": {"status": "no_coverage", "reason": "provider not monitored"}, "policy": {"status": "no_coverage", "reason": "policy not tracked"}, "attestation": {"status": "no_coverage", "effective": "UNKNOWN", "reason": "no attestation"}, "pricing": {"status": "no_coverage"}}}], "summary": {"avoid": 1, "optimal": 3, "insufficient_evidence": 7, "review": 2}}, "provenance": ["/mnt/legion/agents/agent-drift/drift.db", "/home/legion33/agent-temp0/temp0.db", "/mnt/legion/agents/agent-incidents/incidents.db", "/mnt/legion/agents/agent-outages/outages.db", "/home/legion33/ai-status/status.db", "/mnt/legion/agents/agent-policydrift/policydrift.db", "/mnt/legion/agents/agent-attest/attest.db", "/home/legion33/legion-organs/public/verdict.sig.json", "/mnt/legion/agents/agent-pricing-v2/pricing.db"], "confidence": 0.7}
+{
+ "criteria": "trust_score attribué à 6 fournisseurs réels sans URL de source",
+ "file": "evidence/trust_grid.json",
+ "original_bytes": 17820,
+ "original_sha256": "da2faaf8c49980ddb5939f27289997b92e092c6473b508135d770b90fe4a58b2",
+ "reason": "fabricated/unsourced claims about third parties",
+ "referenced_by": [
+  "trust-grid-summary.json"
+ ],
+ "retracted_at": "2026-10-12T13:28:36Z",
+ "retraction_note": "retractions/2026-10-12_evidence-retraction.json",
+ "schema": "phoenix-tombstone/0.1",
+ "signature": {
+  "alg": "Ed25519",
+  "public_key_b64": "ENTr9tUMHipzZIBmrau7pH7+Py6szwCaj/9HKRZp4P8=",
+  "sig": "lED1+f77Al1yde1FPrXj+yKPqEpraJcLNxQ2ImCUWx7iZEhU3GnpeOAXk9ofHuy8lwUjitA7NTDKxw/oZ81RAg==",
+  "signed_message": "PHOENIX_TOMBSTONE_V0.1|3a5a866270a7784f652ff4eedfb80ae80aa643488894e673057db1eb336b8de6"
+ },
+ "status": "RETRACTED"
+}

evidence/wave13_fragility_route.json

@@ -1,262 +1,20 @@
 {
-  "generated_at": "2026-09-25T12:10:27+00:00",
-  "product": "fragility_route",
-  "count": 15,
-  "results": [
-    {
-      "provider": "anthropic",
-      "fragility": 0.552,
-      "fragility_class": "brittle",
-      "vix_30d": 2.666,
-      "cascade_rate": 33.3,
-      "latency_avg_ms": 221.7,
-      "min_input_price": 0.25,
-      "max_context": 200000,
-      "eligibility": "cheap_batch,long_context",
-      "route_score": 2.587,
-      "route_verdict": "avoid",
-      "high_stakes_ok": 0,
-      "reasons": "brittle(vix=2.666,cascade=33.3)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00"
-    },
-    {
-      "provider": "cohere",
-      "fragility": 0.298,
-      "fragility_class": "robust",
-      "vix_30d": 1.456,
-      "cascade_rate": 2.2,
-      "latency_avg_ms": 597.1,
-      "min_input_price": 0.15,
-      "max_context": 128000,
-      "eligibility": "cheap_batch,long_context",
-      "route_score": 1.298,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "aucun signal defavorable mesure",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00"
-    },
-    {
-      "provider": "deepseek",
-      "fragility": null,
-      "fragility_class": "unmeasured",
-      "vix_30d": null,
-      "cascade_rate": null,
-      "latency_avg_ms": null,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 1.95,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing; fragility_unmeasured(0.3 par defaut)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00"
-    },
-    {
-      "provider": "google",
-      "fragility": 0.384,
-      "fragility_class": "watch",
-      "vix_30d": 1.832,
-      "cascade_rate": 3.2,
-      "latency_avg_ms": 495.8,
-      "min_input_price": 0.0375,
-      "max_context": 2000000,
-      "eligibility": "cheap_batch,long_context",
-      "route_score": 0.346,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "aucun signal defavorable mesure",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00"
-    },
-    {
-      "provider": "groq",
-      "fragility": 0.391,
-      "fragility_class": "watch",
-      "vix_30d": 1.456,
-      "cascade_rate": 33.3,
-      "latency_avg_ms": 828.2,
-      "min_input_price": 0.05,
-      "max_context": 131072,
-      "eligibility": "cheap_batch,long_context",
-      "route_score": 0.464,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "slow(828.2ms)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00"
-    },
-    {
-      "provider": "huggingface",
-      "fragility": 0.322,
-      "fragility_class": "watch",
-      "vix_30d": 1.583,
-      "cascade_rate": 1.9,
-      "latency_avg_ms": 591.3,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 1.983,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00"
-    },
-    {
-      "provider": "meta",
-      "fragility": null,
-      "fragility_class": "unmeasured",
-      "vix_30d": null,
-      "cascade_rate": null,
-      "latency_avg_ms": null,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 1.95,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing; fragility_unmeasured(0.3 par defaut)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00"
-    },
-    {
-      "provider": "mistral",
-      "fragility": 0.408,
-      "fragility_class": "watch",
-      "vix_30d": 2.009,
-      "cascade_rate": 2.5,
-      "latency_avg_ms": 416.5,
-      "min_input_price": 0.15,
-      "max_context": 262144,
-      "eligibility": "cheap_batch,long_context",
-      "route_score": 1.408,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "aucun signal defavorable mesure",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00"
-    },
-    {
-      "provider": "nvidia",
-      "fragility": null,
-      "fragility_class": "unmeasured",
-      "vix_30d": null,
-      "cascade_rate": null,
-      "latency_avg_ms": null,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 1.95,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing; fragility_unmeasured(0.3 par defaut)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00"
-    },
-    {
-      "provider": "ollama_hub",
-      "fragility": 0.416,
-      "fragility_class": "watch",
-      "vix_30d": 1.579,
-      "cascade_rate": 33.3,
-      "latency_avg_ms": 678.9,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 2.124,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00"
-    },
-    {
-      "provider": "openai",
-      "fragility": 0.392,
-      "fragility_class": "watch",
-      "vix_30d": 1.635,
-      "cascade_rate": 2.2,
-      "latency_avg_ms": 414.0,
-      "min_input_price": 0.15,
-      "max_context": 200000,
-      "eligibility": "cheap_batch,long_context",
-      "route_score": 1.392,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "aucun signal defavorable mesure",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00"
-    },
-    {
-      "provider": "qwen",
-      "fragility": null,
-      "fragility_class": "unmeasured",
-      "vix_30d": null,
-      "cascade_rate": null,
-      "latency_avg_ms": null,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 1.95,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing; fragility_unmeasured(0.3 par defaut)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00"
-    },
-    {
-      "provider": "together",
-      "fragility": 0.283,
-      "fragility_class": "robust",
-      "vix_30d": 1.248,
-      "cascade_rate": 11.1,
-      "latency_avg_ms": 884.5,
-      "min_input_price": 0.18,
-      "max_context": 131072,
-      "eligibility": "cheap_batch,long_context",
-      "route_score": 1.54,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "slow(884.5ms)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00"
-    },
-    {
-      "provider": "xai",
-      "fragility": null,
-      "fragility_class": "unmeasured",
-      "vix_30d": null,
-      "cascade_rate": null,
-      "latency_avg_ms": null,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 1.95,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing; fragility_unmeasured(0.3 par defaut)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00"
-    },
-    {
-      "provider": "zhipu",
-      "fragility": null,
-      "fragility_class": "unmeasured",
-      "vix_30d": null,
-      "cascade_rate": null,
-      "latency_avg_ms": null,
-      "min_input_price": null,
-      "max_context": null,
-      "eligibility": null,
-      "route_score": 1.95,
-      "route_verdict": "route",
-      "high_stakes_ok": 0,
-      "reasons": "no_pricing; fragility_unmeasured(0.3 par defaut)",
-      "method": "route_score = (prix/median) x (1+fragilite) | fragilite inconnue = 0.3 declaree",
-      "computed_at": "2026-09-25T12:10:26+00:00"
-    }
-  ]
-}
+ "criteria": "scores de fragilité/route attribués à 12 fournisseurs réels sans URL de source",
+ "file": "evidence/wave13_fragility_route.json",
+ "original_bytes": 8661,
+ "original_sha256": "dc24b7602aa31c3b9e5fedacb9846d0887497b1096fdb2ca98c81c0fbaa44a99",
+ "reason": "fabricated/unsourced claims about third parties",
+ "referenced_by": [
+  "snapshot-wave13_fragility_route.json"
+ ],
+ "retracted_at": "2026-10-12T13:28:36Z",
+ "retraction_note": "retractions/2026-10-12_evidence-retraction.json",
+ "schema": "phoenix-tombstone/0.1",
+ "signature": {
+  "alg": "Ed25519",
+  "public_key_b64": "ENTr9tUMHipzZIBmrau7pH7+Py6szwCaj/9HKRZp4P8=",
+  "sig": "WcxSEkIn7q4VDa6etw6gUFBWhoerQ2P4vFYwNQdfeE/5IhwDCxwlW+czwms0dauhInvqwNAgukIEvLv8NZLvDg==",
+  "signed_message": "PHOENIX_TOMBSTONE_V0.1|65dee76bd474f85aa5215bbeab7161ea33687e83376dd8bea89d87973d36cb9f"
+ },
+ "status": "RETRACTED"
+}

phoenix_verify.py

@@ -1,11 +1,24 @@
 #!/usr/bin/env python3
 # PHOENIX verifier — autonome, aucune dependance a LEGION. Ne plante jamais.
-# Verdicts: VALID_FULL (signature + toutes preuves) | VALID_SIGNATURE_ONLY (preuves non fournies) | INVALID
+# Verdicts: VALID_FULL (signature + toutes preuves) | VALID_SIGNATURE_ONLY (preuves non fournies ou RETIREES par tombstone signe) | INVALID
 # Bundle: python3 phoenix_verify.py     Unitaire: python3 phoenix_verify.py objet.json cle.json [dossier_preuves]
 # Code retour: 0 tout VALID_FULL | 3 au moins un VALID_SIGNATURE_ONLY | 1 au moins un INVALID | 2 usage
 import sys, os, json, glob, hashlib, base64
 F = ("phoenix", "subject", "observation", "claim", "evidence")
 def canon(o): return json.dumps({k: o[k] for k in F}, sort_keys=True, separators=(",", ":"), ensure_ascii=False).encode()
+def tombstone_ok(p, expected_sha256, pub_b64):
+    """Pièce retirée : tombstone phoenix-tombstone/0.1 signé par la clé de l'objet et portant l'empreinte de l'original."""
+    try:
+        t = json.load(open(p))
+        sg = t.get("signature") or {}
+        body = {k: v for k, v in t.items() if k != "signature"}
+        d = hashlib.sha256(json.dumps(body, sort_keys=True, separators=(",", ":"), ensure_ascii=False).encode()).hexdigest()
+        if not (t.get("schema") == "phoenix-tombstone/0.1" and t.get("status") == "RETRACTED" and t.get("original_sha256") == expected_sha256
+                and sg.get("public_key_b64") == pub_b64 and sg.get("signed_message") == "PHOENIX_TOMBSTONE_V0.1|" + d): return False
+        from cryptography.hazmat.primitives.asymmetric.ed25519 import Ed25519PublicKey
+        Ed25519PublicKey.from_public_bytes(base64.b64decode(pub_b64)).verify(base64.b64decode(sg.get("sig", "")), sg["signed_message"].encode())
+        return True
+    except Exception: return False
 def find_ev(edir, uri):
     for p in (os.path.join(edir, uri), os.path.join(edir, os.path.basename(uri))):
         if os.path.isfile(p): return p
@@ -34,6 +47,8 @@ def check(path, key, edir):
             p = find_ev(edir, str(e.get("uri", ""))) if edir else None
             if not p: missing += 1; print("  preuve non fournie:", e.get("uri")); continue
             ok = hashlib.sha256(open(p, "rb").read()).hexdigest() == e.get("sha256")
+            if not ok and tombstone_ok(p, e.get("sha256"), key.get("key")):
+                missing += 1; print("  preuve %s RETIREE (tombstone signe, voir retractions/)" % os.path.basename(p)); continue
             print("  preuve %s %s" % (os.path.basename(p), "OK" if ok else "ALTEREE"))
             if not ok: fail.append("preuve alteree: " + os.path.basename(p))
         if fail: return "INVALID", fail

phoenix_verify2.py

@@ -54,6 +54,19 @@ def load_keyring(path):
             else: probs.append("type inconnu ligne %d" % i)
         except Exception as x: probs.append("ligne %d: %s" % (i, type(x).__name__))
     return keys, probs
+def tombstone_ok(p, expected_sha256, pub_b64):
+    """Pièce retirée : tombstone phoenix-tombstone/0.1 signé par la clé de l'objet et portant l'empreinte de l'original."""
+    try:
+        t = json.load(open(p))
+        sg = t.get("signature") or {}
+        body = {k: v for k, v in t.items() if k != "signature"}
+        d = hashlib.sha256(json.dumps(body, sort_keys=True, separators=(",", ":"), ensure_ascii=False).encode()).hexdigest()
+        if not (t.get("schema") == "phoenix-tombstone/0.1" and t.get("status") == "RETRACTED" and t.get("original_sha256") == expected_sha256
+                and sg.get("public_key_b64") == pub_b64 and sg.get("signed_message") == "PHOENIX_TOMBSTONE_V0.1|" + d): return False
+        from cryptography.hazmat.primitives.asymmetric.ed25519 import Ed25519PublicKey
+        Ed25519PublicKey.from_public_bytes(base64.b64decode(pub_b64)).verify(base64.b64decode(sg.get("sig", "")), sg["signed_message"].encode())
+        return True
+    except Exception: return False
 def check(path, edir, keys, pinset, pin):
     flags, fail = [], []
     try:
@@ -65,7 +78,7 @@ def check(path, edir, keys, pinset, pin):
         d = hashlib.sha256(canon(obj)).hexdigest(); pk = str(att.get("public_key_b64", ""))
         (flags.append("CONTENT_INTACT") if att.get("digest") == d else fail.append("contenu modifie"))
         (flags.append("SIGNATURE_VALID") if vsig(pk, att.get("signature", ""), "PHOENIX_V0.1|" + d) else fail.append("signature invalide"))
-        miss = 0
+        miss = retr = 0
         for e in ev:
             p = None
             if isinstance(e, dict) and edir:
@@ -73,8 +86,10 @@ def check(path, edir, keys, pinset, pin):
                 for q in (os.path.join(edir, u), os.path.join(edir, os.path.basename(u))):
                     if os.path.isfile(q): p = q; break
             if not p: miss += 1; continue
-            if hashlib.sha256(open(p, "rb").read()).hexdigest() != e.get("sha256"): fail.append("preuve alteree: " + os.path.basename(p))
-        flags.append("EVIDENCE_MISSING" if miss else "EVIDENCE_ALTERED" if any(f.startswith("preuve") for f in fail) else "EVIDENCE_INTACT")
+            if hashlib.sha256(open(p, "rb").read()).hexdigest() != e.get("sha256"):
+                if tombstone_ok(p, e.get("sha256"), pk): retr += 1; continue
+                fail.append("preuve alteree: " + os.path.basename(p))
+        flags.append("EVIDENCE_MISSING" if miss else "EVIDENCE_ALTERED" if any(f.startswith("preuve") for f in fail) else "EVIDENCE_RETRACTED" if retr else "EVIDENCE_INTACT")
         try: k, f = kid(pk), fp(pk)
         except Exception: k = f = None
         flags.append("ISSUER_KEY_PINNED" if k and (k in pinset or (pin and f == pin)) else "ISSUER_NOT_PINNED")
@@ -109,8 +124,8 @@ def main():
         print("  ->", " ".join(fl) + ("" if not fa else " | INVALID: " + "; ".join(fa)))
         if fa or "INVALID" in fl: n["invalid"] += 1
         elif "ISSUER_NOT_PINNED" in fl: n["unpinned"] += 1
-        elif "EVIDENCE_MISSING" in fl: n["missing"] += 1
-    print("\nBILAN: %d objet(s) | invalides %d | emetteur non epingle %d | preuves manquantes %d%s" % (len(objs), n["invalid"], n["unpinned"], n["missing"], "" if pin else " (aucun --pin)"))
+        elif "EVIDENCE_MISSING" in fl or "EVIDENCE_RETRACTED" in fl: n["missing"] += 1
+    print("\nBILAN: %d objet(s) | invalides %d | emetteur non epingle %d | preuves manquantes ou retirees %d%s" % (len(objs), n["invalid"], n["unpinned"], n["missing"], "" if pin else " (aucun --pin)"))
     if kr_bad: print("KEYRING REJETE: une revocation a pu etre effacee; resultat non fiable (rc=5)"); return 5
     return 1 if n["invalid"] else 4 if n["unpinned"] else 3 if n["missing"] else 0
 if __name__ == "__main__": sys.exit(main())

retractions/2026-10-12_evidence-retraction.json

@@ -0,0 +1,108 @@
+{
+ "criteria": [
+  "attributes a score, risk, amount or premium to a named real entity without a source URL",
+  "carries \"measured\": true without a source"
+ ],
+ "files": [
+  {
+   "criteria": "scores/montants attribués à des entités réelles nommées sans URL de source (11 enregistrements) ; \"measured\": true sans source ; réutilise le chiffre non traçable « 1245 incidents/90d »",
+   "file": "evidence/signals.json",
+   "original_bytes": 18582,
+   "original_sha256": "f558b714ad1d42958509c46de0c435a596ca4fada9476c20fe78b7236a66230a",
+   "referenced_by": [
+    "snapshot-signals.json"
+   ]
+  },
+  {
+   "criteria": "primes d'assurance (USD), niveaux et scores composites attribués à 7 fournisseurs réels, sans URL de source ; 2 composantes sur 3 déclarées « estimated »",
+   "file": "evidence/risk_matrix.json",
+   "original_bytes": 1395,
+   "original_sha256": "c97174932910256a346de49edf600a40907541878ec22bada7b740cab01ec3d9",
+   "referenced_by": [
+    "snapshot-risk_matrix.json"
+   ]
+  },
+  {
+   "criteria": "trust_score attribué à 6 fournisseurs réels sans URL de source",
+   "file": "evidence/trust_grid.json",
+   "original_bytes": 17820,
+   "original_sha256": "da2faaf8c49980ddb5939f27289997b92e092c6473b508135d770b90fe4a58b2",
+   "referenced_by": [
+    "trust-grid-summary.json"
+   ]
+  },
+  {
+   "criteria": "scores de confiance attribués à des fournisseurs réels sans URL de source (13 enregistrements)",
+   "file": "evidence/negative_space.json",
+   "original_bytes": 166152,
+   "original_sha256": "a8e39c5b4c71cbd2246ce8f6752b3ba4749149c04c38aa2fdff4139acc4e8097",
+   "referenced_by": [
+    "snapshot-negative_space.json"
+   ]
+  },
+  {
+   "criteria": "scores attribués à des entités réelles nommées, dérivés de bases internes, sans URL de source",
+   "file": "evidence/entities.json",
+   "original_bytes": 79025,
+   "original_sha256": "52be94dfef707193f006b1722e3f8ce88afb6d0a7be5cae710969c6ae0f12e04",
+   "referenced_by": [
+    "snapshot-entities.json"
+   ]
+  },
+  {
+   "criteria": "scores de fragilité/route attribués à 12 fournisseurs réels sans URL de source",
+   "file": "evidence/fragility_confidence.json",
+   "original_bytes": 11023,
+   "original_sha256": "dc0b6481fede7a3efac6cc3683f1e73c6c6f11210633eec94a17555d63dba565",
+   "referenced_by": [
+    "snapshot-fragility_confidence.json"
+   ]
+  },
+  {
+   "criteria": "trust_score attribué à 6 fournisseurs réels sans URL de source",
+   "file": "evidence/eligibility.json",
+   "original_bytes": 5555,
+   "original_sha256": "5ccfce2f97d9e73fe6bd92f02d2ae03eed1f375ec223d3e5cfccee6a64a94091",
+   "referenced_by": [
+    "snapshot-eligibility.json"
+   ]
+  },
+  {
+   "criteria": "scores de fragilité/route attribués à 12 fournisseurs réels sans URL de source",
+   "file": "evidence/wave13_fragility_route.json",
+   "original_bytes": 8661,
+   "original_sha256": "dc24b7602aa31c3b9e5fedacb9846d0887497b1096fdb2ca98c81c0fbaa44a99",
+   "referenced_by": [
+    "snapshot-wave13_fragility_route.json"
+   ]
+  },
+  {
+   "criteria": "15 enregistrements \"measured\": true sans aucune source (dépôts GitHub tiers classés « ADOPT_CANDIDATE »)",
+   "file": "evidence/orphan_adoption.json",
+   "original_bytes": 4979,
+   "original_sha256": "479928f23455d23e5e7c7a7d16a8dec745ebc221b8079778416526fe638dfa30",
+   "referenced_by": [
+    "snapshot-orphan_adoption.json"
+   ]
+  }
+ ],
+ "how_to_check": "python3 phoenix_verify.py ; python3 phoenix_verify2.py --keyring keyring.jsonl --pin a7c6a35b1c2c81d589a30c2ad628679bdea9da4139d49aa3c42cdde6e1c65036",
+ "issuer": "LEGION (https://legion-api.com)",
+ "not_affected": [
+  "tlog/ (transparency-log mirror, export 15192-r2) is untouched",
+  "objects/ are untouched (63 signed objects)",
+  "keyring.jsonl and legion-receipt-key.json are untouched"
+ ],
+ "originals": "kept offline by LEGION; each original is identified by original_sha256",
+ "reason": "fabricated/unsourced claims about third parties",
+ "retracted_at": "2026-10-12T13:28:36Z",
+ "schema": "phoenix-retraction/0.1",
+ "signature": {
+  "alg": "Ed25519",
+  "public_key_b64": "ENTr9tUMHipzZIBmrau7pH7+Py6szwCaj/9HKRZp4P8=",
+  "sig": "zzi1FveBrwPECDHAi4JGmkRbmbtj7gC6+535GTFlB+e42Cwvdo8tXCuo0F3C53yTBwGDD2yc43Ci0B4MEEsRAA==",
+  "signed_message": "PHOENIX_RETRACTION_V0.1|631e64771668bb487424b9d5da3e2a429734e4e4e7a1188fdb9e96f193a2a1a8"
+ },
+ "statement": "LEGION retracts the evidence files listed below. They contained scores, risk tiers, monetary amounts or premiums attributed to named real companies without any source URL, or values labelled \"measured\": true without any source. They were published here on 2026-10-03 and are not deleted silently: each file is replaced by a signed tombstone carrying the SHA-256 of the original, so their past existence stays provable. The Phoenix objects that reference them are unchanged; their signatures stay valid, but their evidence is now RETRACTED (verdict VALID_SIGNATURE_ONLY, never VALID_FULL).",
+ "title": "Retraction of fabricated / unsourced evidence files in phoenix-objects"
+}