Add three attributed ASTRIDE evaluation illustrations

CITATION.bib

@@ -3,5 +3,14 @@
   year = {2026},
   howpublished = {Hugging Face Datasets},
   url = {https://huggingface.co/datasets/lambdaeranga/astride},
-  note = {Version v0.1.0}
+  note = {Version v0.2.0}
+}
+
+@misc{bandara2025astride,
+  title = {{ASTRIDE}: A Security Threat Modeling Platform for Agentic-{AI} Applications},
+  author = {Bandara, Eranga and Hass, Amin and Gore, Ross and Shetty, Sachin and Mukkamala, Ravi and Bouk, Safdar H. and Liang, Xueping and Keong, Ng Wee and De Zoysa, Kasun and Withanage, Aruna and Loganathan, Nilaan},
+  year = {2025},
+  howpublished = {arXiv:2512.04785},
+  doi = {10.48550/arXiv.2512.04785},
+  url = {https://arxiv.org/abs/2512.04785}
 }

README.md

@@ -28,11 +28,17 @@ configs:
   data_files:
   - split: evaluation
     path: data/trials/evaluation-00000-of-00001.parquet
+- config_name: illustrations
+  data_files:
+  - split: examples
+    path: data/illustrations/examples-00000-of-00001.parquet
 ---
 
 # ASTRIDE: Diagram-Grounded Threat-Modeling Dataset
 
-Version **v0.1.0** contains synthetic architecture diagrams, operating facts, candidate threat annotations, mitigation recommendations, and generation trial outcomes. It supports research on constructing datasets for ASTRIDE-style vision–language threat analysis.
+Version **v0.2.0** adds three attributed Mermaid redraws of architectures from the original ASTRIDE evaluation in a separate `illustrations` subset. The original annotated collections and trial outcomes are unchanged from `v0.1.0`.
+
+The release contains synthetic architecture diagrams, operating facts, candidate threat annotations, mitigation recommendations, and generation trial outcomes. It supports research on constructing datasets for ASTRIDE-style vision–language threat analysis.
 
 The dataset accompanies the draft **An Agentic Framework for Constructing and Evaluating Multimodal Threat-Modeling Datasets**. It is distinct from the corpus used in the original ASTRIDE threat-analysis paper. A publication reference will be added when available.
 
@@ -43,8 +49,9 @@ The dataset accompanies the draft **An Agentic Framework for Constructing and Ev
 | `comparison` (default) | `evaluation` | 48 diagrams | 154 |
 | `development` | `examples` | 3 diagrams | 14 |
 | `trials` | `evaluation` | 54 generation attempts | Not applicable |
+| `illustrations` | `examples` | 3 architecture redraws | Not supplied |
 
-There are **51 diagram records and 168 candidate threats** in total. The controlled experiment uses only the 48 comparison diagrams and 154 threats. The three earlier development examples used a different prompt and recovery history and must not be pooled into that experiment. The 36-record authored template pilot is not included. No training split or verified threat-detection benchmark is provided.
+There are **51 annotated diagram records and 168 candidate threats**, plus **3 architecture illustrations** (54 diagram images total). The controlled experiment uses only the 48 comparison diagrams and 154 threats. The three earlier development examples used a different prompt and recovery history and must not be pooled into that experiment. The 36-record authored template pilot is not included. No training split or verified threat-detection benchmark is provided.
 
 ## Load diagrams and annotations
 
@@ -52,7 +59,7 @@ There are **51 diagram records and 168 candidate threats** in total. The control
 from datasets import load_dataset
 
 repo = "lambdaeranga/astride"
-ds = load_dataset(repo, "comparison", split="evaluation", revision="v0.1.0")
+ds = load_dataset(repo, "comparison", split="evaluation", revision="v0.2.0")
 example = ds[0]
 image = example["image"]  # PIL image; PNG bytes are embedded in Parquet
 facts = {k: example[k] for k in ("title", "purpose", "assumptions", "controls")}
@@ -60,15 +67,35 @@ architecture = example["architecture"]
 threats = example["threats"]
 mitigations = example["mitigations"]
 
-examples = load_dataset(repo, "development", split="examples", revision="v0.1.0")
-trials = load_dataset(repo, "trials", split="evaluation", revision="v0.1.0")
+examples = load_dataset(repo, "development", split="examples", revision="v0.2.0")
+trials = load_dataset(repo, "trials", split="evaluation", revision="v0.2.0")
 ```
 
-Images load without any files from the author's workstation. Every diagram is also provided as a PNG, SVG, and Mermaid source under `images/`. Portable JSONL records under `records/` use image paths relative to this repository.
+Images load without any files from the author's workstation. Every diagram is also provided as a PNG, SVG, and Mermaid source under `images/`. The three `illustrations` rows provide source attribution and architecture images without threat/mitigation labels. Portable JSONL records under `records/` use image paths relative to this repository.
 
 For VLM input, use the **image plus operating facts**. Canonical architecture labels, threats, mitigations, and generation-condition identifiers are targets or study metadata; do not leak them into the prediction input.
 
-## Record fields
+## ASTRIDE evaluation illustrations
+
+The `illustrations` subset accompanies the downstream evaluation discussion in the dataset-generation paper. These are Mermaid redraws of architecture panels from [ASTRIDE: A Security Threat Modeling Platform for Agentic-AI Applications](https://arxiv.org/abs/2512.04785), Figures 7–9.
+
+| Record | Source figure | Dataset-generation paper figure | Evaluation context |
+|---|---:|---:|---|
+| `astride-llama-architecture` | 7 | 8 | Tool-using agent; Llama-Vision before/after example |
+| `astride-pixtral-architecture` | 8 | 9 | Voice-controlled agent; Pixtral-Vision before/after example |
+| `astride-synthesis-architecture` | 9 | 10 | Multi-agent collaboration; VLM predictions and reasoning synthesis |
+
+```python
+illustrations = load_dataset(repo, "illustrations", split="examples", revision="v0.2.0")
+image = illustrations[0]["image"]
+source = illustrations[0]["source_paper"]
+```
+
+The redraws preserve source components, groups, and visible directed connections. They add reference IDs, normalize labels, standardize storage symbols, and revise layout. They are **illustrations, not new agent-generated examples, training targets, or additional experimental results**. Their `annotation_status` is `illustration_only`. No complete threat/mitigation annotation is supplied, and these rows must not be pooled into the frozen generation comparison. The prior ASTRIDE training corpus is not included in this release.
+
+Illustration fields include `id`, `title`, embedded `image`, PNG/SVG/Mermaid paths, `source_paper`, `source_version`, `source_figure`, `manuscript_figure`, `evaluation_context`, `origin`, `annotation_status`, `adaptation`, `scope`, and `mermaid_sha256`. Cite both the dataset release and the original ASTRIDE paper when using these redraws. Source code, prompts, raw traces, manuscript PDFs, and detailed ratings remain excluded.
+
+## Annotated record fields
 
 - `id`, `collection`, `trial_id`: stable record and trial identifiers.
 - `domain`, `control_profile`, `generation_condition`: experiment grouping metadata.
@@ -109,13 +136,16 @@ data/
   comparison/evaluation-00000-of-00001.parquet
   development/examples-00000-of-00001.parquet
   trials/evaluation-00000-of-00001.parquet
+  illustrations/examples-00000-of-00001.parquet
 records/
   comparison.jsonl
   development.jsonl
   trials.jsonl
+  illustrations.jsonl
 images/
   comparison/*.{png,svg,mmd}
   development/*.{png,svg,mmd}
+  illustrations/*.{png,svg,mmd}
 ```
 
 The release intentionally excludes implementation source, resolved generation prompts, raw model-call traces, intermediate drafts, and detailed evaluation ratings/packets. Protocol and rating hashes in `release.json` are lineage identifiers; the corresponding private artifacts are not accessible through this release. Full experimental reproduction is therefore not supported by this package alone.
@@ -124,7 +154,9 @@ The release intentionally excludes implementation source, resolved generation pr
 
 Use these candidates to inspect synthetic diagram/annotation structure, develop annotation checks, and design prospective VLM training experiments. Known defects include unsupported threats, overstated certainty, partial mitigations, and questionable graph semantics. These annotations must not be treated as verified vulnerabilities or automatically correct training targets. Suggested verification procedures were not executed. The diagrams do not represent actual customers, payments, or incident records.
 
-Some diagrams share typed topology. There are 44 typed-isomorphism groups across the 51 diagrams, and semantic duplication can remain between different groups. Keep related graph families and all derived renderings together in any future split. The current `evaluation` split identifies a generation-study collection, not an independently validated benchmark.
+Some diagrams share typed topology. There are 44 typed-isomorphism groups across the 51 annotated diagrams; the three illustration redraws are outside this grouping analysis. Semantic duplication can remain between different groups. Keep related graph families and their renderings together in future splits.
+
+The current `evaluation` split identifies a generation-study collection, not an independently validated benchmark.
 
 ## License and citation
 
@@ -138,6 +170,6 @@ Use `CITATION.bib` and identify the exact release tag or Hub revision used:
   year = {2026},
   howpublished = {Hugging Face Datasets},
   url = {https://huggingface.co/datasets/lambdaeranga/astride},
-  note = {Version v0.1.0}
+  note = {Version v0.2.0}
 }
 ```

data/illustrations/examples-00000-of-00001.parquet

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f7e1743f6c00dd4481c6d34735df7a2faa0d2edda8df690c6ea1139dc3dc6017
+size 134913

images/illustrations/astride-llama-architecture.mmd

@@ -0,0 +1,27 @@
+%% Redrawn from ASTRIDE Figure 7. IDs are added for this manuscript.
+flowchart TB
+    subgraph USER["User"]
+        E1["E1: End user"]
+    end
+    subgraph AGENT["AI agent"]
+        P1("P1: User<br/>interface")
+        P2("P2: Prompt<br/>processor")
+        A1("A1: LLM<br/>reasoning core")
+        P3("P3: Tool<br/>execution module")
+    end
+    subgraph EXTERNAL["External APIs"]
+        E2["E2: Web<br/>search API"]
+        E3["E3: Payment<br/>gateway API"]
+    end
+    subgraph INTERNAL["Internal"]
+        P4("P4: Audit<br/>and logging")
+        D1[("D1: Vector<br/>memory store")]
+    end
+    E1 --> P1
+    P1 --> P2
+    P2 --> A1
+    A1 --> P3
+    A1 --> D1
+    P3 --> E2
+    P3 --> E3
+    P3 --> P4

images/illustrations/astride-pixtral-architecture.mmd

@@ -0,0 +1,29 @@
+%% Redrawn from ASTRIDE Figure 8. IDs are added for this manuscript.
+flowchart TB
+    subgraph USER["User"]
+        E1["E1: User<br/>voice input"]
+    end
+    subgraph AGENT["AI agent"]
+        P1("P1: Speech-to-text<br/>engine")
+        A1("A1: LLM-based<br/>understanding")
+        A2("A2: Intent and<br/>reasoning module")
+        P2("P2: Voice<br/>tool executor")
+    end
+    subgraph INTERNAL["Internal"]
+        D1[("D1: Contextual<br/>memory")]
+        D2[("D2: Interaction<br/>logs")]
+    end
+    subgraph THIRD["Third party"]
+        E2["E2: Smart<br/>home API"]
+        E3["E3: Calendar<br/>integration"]
+    end
+    E1 --> P1
+    P1 --> A1
+    A1 --> A2
+    A2 --> D1
+    A2 --> D2
+    A2 --> P2
+    P2 --> E2
+    P2 --> E3
+    %% Invisible spacing constraint; not an architectural flow.
+    D1 ~~~ D2

images/illustrations/astride-synthesis-architecture.mmd

@@ -0,0 +1,28 @@
+%% Redrawn from ASTRIDE Figure 9. IDs are added for this manuscript.
+flowchart TB
+    subgraph USER["User"]
+        E1["E1: Human<br/>analyst"]
+    end
+    subgraph AGENTS["Agents"]
+        A1("A1: Planning<br/>agent")
+        A2("A2: Execution<br/>agent")
+        A3("A3: Review<br/>agent")
+    end
+    subgraph BACKEND["Backend"]
+        D1[("D1: Task<br/>queue")]
+        D2[("D2: Shared<br/>memory store")]
+    end
+    subgraph TOOLS["External tools"]
+        P1("P1: Internal<br/>API service")
+        E2["E2: External<br/>web API"]
+    end
+    E1 --> A1
+    A1 --> D1
+    D1 --> A2
+    A2 --> D2
+    D2 --> A3
+    A3 --> E1
+    A2 --> P1
+    A2 --> E2
+    %% Invisible spacing constraint; not an architectural flow.
+    P1 ~~~ E2

records/illustrations.jsonl

@@ -0,0 +1,3 @@
+{"id": "astride-llama-architecture", "collection": "illustrations", "title": "Tool-using AI agent", "image_path": "images/illustrations/astride-llama-architecture.png", "svg_path": "images/illustrations/astride-llama-architecture.svg", "mermaid_path": "images/illustrations/astride-llama-architecture.mmd", "source_paper": "https://arxiv.org/abs/2512.04785", "source_version": "v1", "source_figure": 7, "manuscript_figure": 8, "evaluation_context": "Llama-Vision before/after fine-tuning example", "origin": "mermaid_redraw_of_prior_paper", "annotation_status": "illustration_only", "mermaid_sha256": "dfb11fd4289b306c36594d951ca3e0aa3ffcd33a74ca894883dd14630c26ebb9", "adaptation": "Reference IDs added; labels normalized; storage symbols standardized; layout revised. Source components, groups, and visible directed connections retained.", "scope": "Architecture illustration accompanying reported ASTRIDE results. No new generation trial, prediction, complete threat annotation, or mitigation measurement is supplied."}
+{"id": "astride-pixtral-architecture", "collection": "illustrations", "title": "Voice-controlled AI agent", "image_path": "images/illustrations/astride-pixtral-architecture.png", "svg_path": "images/illustrations/astride-pixtral-architecture.svg", "mermaid_path": "images/illustrations/astride-pixtral-architecture.mmd", "source_paper": "https://arxiv.org/abs/2512.04785", "source_version": "v1", "source_figure": 8, "manuscript_figure": 9, "evaluation_context": "Pixtral-Vision before/after fine-tuning example", "origin": "mermaid_redraw_of_prior_paper", "annotation_status": "illustration_only", "mermaid_sha256": "0e22f4d4d69696413c83edf6e431f6ffc11f09a7a938f74ec29b48412624c196", "adaptation": "Reference IDs added; labels normalized; storage symbols standardized; layout revised. Source components, groups, and visible directed connections retained.", "scope": "Architecture illustration accompanying reported ASTRIDE results. No new generation trial, prediction, complete threat annotation, or mitigation measurement is supplied."}
+{"id": "astride-synthesis-architecture", "collection": "illustrations", "title": "Multi-agent collaboration", "image_path": "images/illustrations/astride-synthesis-architecture.png", "svg_path": "images/illustrations/astride-synthesis-architecture.svg", "mermaid_path": "images/illustrations/astride-synthesis-architecture.mmd", "source_paper": "https://arxiv.org/abs/2512.04785", "source_version": "v1", "source_figure": 9, "manuscript_figure": 10, "evaluation_context": "VLM predictions and OpenAI-gpt-oss synthesis example", "origin": "mermaid_redraw_of_prior_paper", "annotation_status": "illustration_only", "mermaid_sha256": "c7429ccb34c9509d7040ed8476c1281fa2f0a797f08d079f2fa18c05c8e693a3", "adaptation": "Reference IDs added; labels normalized; storage symbols standardized; layout revised. Source components, groups, and visible directed connections retained.", "scope": "Architecture illustration accompanying reported ASTRIDE results. No new generation trial, prediction, complete threat annotation, or mitigation measurement is supplied."}

release.json

@@ -1,5 +1,5 @@
 {
-  "version": "v0.1.0",
+  "version": "v0.2.0",
   "proposed_repo_id": "lambdaeranga/astride",
   "license": null,
   "package_status": "dataset_release",
@@ -17,6 +17,11 @@
     "trials": {
       "records": 54,
       "split": "evaluation"
+    },
+    "illustrations": {
+      "records": 3,
+      "split": "examples",
+      "annotation_status": "illustration_only"
     }
   },
   "topology_families": 44,
@@ -31,5 +36,7 @@
     "intermediate_drafts",
     "evaluation_ratings",
     "evaluation_packets"
-  ]
+  ],
+  "previous_version": "v0.1.0",
+  "topology_families_scope": "Original 51 annotated records only; illustration redraws are not assigned topology families."
 }

release_manifest.json

@@ -1,9 +1,10 @@
 {
   "files": {
-    "CITATION.bib": "8e7e3291f28151f36d13e999e89ae92796376575974ddff659edbecb1e832637",
-    "README.md": "ec4d3314ab81c276f54ad6b12e28075301ffe32944dd8df649424a8cbd71979b",
+    "CITATION.bib": "81698b44e4398f6ed5287e5564a843cc9a7804ce674e3132601981d49ebc6607",
+    "README.md": "b42fe10aadaf889cfa0b79af0c60e8f4aeb0a39d2b80c56c9acb06833f692729",
     "data/comparison/evaluation-00000-of-00001.parquet": "ca163b1aa6377f6733b46a69b0bd47acb9eb55938192ceb70a10f21966bf4151",
     "data/development/examples-00000-of-00001.parquet": "254af98e07912e2646f0a7797a56714f217de6607454a5a00016a90aeee9fb00",
+    "data/illustrations/examples-00000-of-00001.parquet": "f7e1743f6c00dd4481c6d34735df7a2faa0d2edda8df690c6ea1139dc3dc6017",
     "data/trials/evaluation-00000-of-00001.parquet": "8f7400ab3be5fcda2a2da4b14c96652d82c741d4d50f44b77fccbc761aa3a148",
     "images/comparison/E001.mmd": "b114af78de457b854e518f6c2682ef0fb871b323f15acc25017f12baf8c08d66",
     "images/comparison/E001.png": "ffdba1ed0783e7b77fc310b80a820687677780445b6f5b971f78ec37646b61de",
@@ -158,10 +159,20 @@
     "images/development/DEV-diagram-0003.mmd": "015946d4ac7de0821a42889a1933fc16d754471ebb485241fc90c4128dedc4dd",
     "images/development/DEV-diagram-0003.png": "0b8322f85ce76c70ecd5d0c4879476848a042c25f7efdd5a1ced83e941fd9605",
     "images/development/DEV-diagram-0003.svg": "2610ea04515984dd296341b0e43903d88d7abe47bbe994ffd690c71cbe8a4a53",
+    "images/illustrations/astride-llama-architecture.mmd": "dfb11fd4289b306c36594d951ca3e0aa3ffcd33a74ca894883dd14630c26ebb9",
+    "images/illustrations/astride-llama-architecture.png": "7f2b9d96ed0ea68d891a576ba4506585a540d0b05122e9e0bb34bf4386673479",
+    "images/illustrations/astride-llama-architecture.svg": "b62eff6263a0f5147bf1a583cfdd633f787edc4080d31463e99440e361bd6652",
+    "images/illustrations/astride-pixtral-architecture.mmd": "0e22f4d4d69696413c83edf6e431f6ffc11f09a7a938f74ec29b48412624c196",
+    "images/illustrations/astride-pixtral-architecture.png": "04591e9c2753957ca6d6d9274b5b318a73ca92f30341357fa79c950f43aa0717",
+    "images/illustrations/astride-pixtral-architecture.svg": "c20f121d16e3fc67f997de85e8ee0918b089f7a64f74f86b7b3e472a89eaa213",
+    "images/illustrations/astride-synthesis-architecture.mmd": "c7429ccb34c9509d7040ed8476c1281fa2f0a797f08d079f2fa18c05c8e693a3",
+    "images/illustrations/astride-synthesis-architecture.png": "18720ac623ec2bf60dd3569d6119050012a3e48b057d8011a8a1b619cb016cad",
+    "images/illustrations/astride-synthesis-architecture.svg": "179e6180dc52f56c51e889e16c0d98f5d7ddd8749ed6097110fa2bec96184e59",
     "records/comparison.jsonl": "f01debc3cd3c2e0d9df68d03ed7941906a1d6ad0437d68158b0024e2d5b3e76b",
     "records/development.jsonl": "68bd2a0eb034267576642e601019e5bb2f17e484979abdc8d10cc5bb793b68f2",
+    "records/illustrations.jsonl": "54795d54fe6ca823b7499ddae2edd9279e7a50151c57676b043b4cc7f030b5a8",
     "records/trials.jsonl": "61b4b6216346245739e9e00feaca3b3fd78863315b8275df8fce2911d909bffd",
-    "release.json": "fdd4b13e9b4f85cd1e3c9aa2e81009bc19791b4c0a4ae215bb4a70d1c99b670c",
-    "schema.json": "621450e601627d575011e17c324fd876507fc1a173013651dc9fa16a5e5c6059"
+    "release.json": "7132640b6c263f341b29cd31c0364cce3aeb8fac43c511eda7d0fd039264e844",
+    "schema.json": "e086a8577e335a8f135c37e3449228a73309e9cde95155fd97da9b1733171e79"
   }
 }

schema.json

@@ -620,5 +620,74 @@
       "dtype": "float64",
       "_type": "Value"
     }
+  },
+  "illustrations": {
+    "id": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "collection": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "title": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "image_path": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "svg_path": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "mermaid_path": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "source_paper": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "source_version": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "source_figure": {
+      "dtype": "int64",
+      "_type": "Value"
+    },
+    "manuscript_figure": {
+      "dtype": "int64",
+      "_type": "Value"
+    },
+    "evaluation_context": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "origin": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "annotation_status": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "mermaid_sha256": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "adaptation": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "scope": {
+      "dtype": "string",
+      "_type": "Value"
+    },
+    "image": {
+      "_type": "Image"
+    }
   }
 }