publication.json

{
 "kind": "sagi.skill_publication/1",
 "loop": "savante",
 "improve_skill_version": "0.0.34",
 "iteration": 34,
 "road": 1000,
 "remaining": 966,
 "memory_lines": 35,
 "refused": 1,
 "skill": {
  "source": "github.com/cryptoAGI/savante",
  "canon_commit": "029642b",
  "sha256": "f5147a9bcc978251964121f1e55b0f635ebb5883a97c2108a435f4b4b1612169",
  "bytes": 11907
 },
 "policy": "auto count, gated text (operator, 2026-09-26)",
 "headroom": {
  "last": 207,
  "mean_last_20": 171.2
 },
 "last": {
  "n": 34,
  "version": "0.0.34",
  "date": "2026-10-12",
  "limit": 257,
  "completion_tokens": 50,
  "headroom": 207
 },
 "derived_by": "scripts/savante_loop.py with the canon's bind/skill_history.py",
 "proposals_queued": []
}