feat: add scoped Spec Studio audits and confirmed remediation

.claude-plugin/skill-assets.sha256

@@ -3,4 +3,4 @@ c5d0c26f28c9ee14092f9deaf24c98dd8bef49d971fef2b7a537ffb1ab9f2887  .claude-plugin
 aeee7a94671ceb306fe2d24c5acc9f2d96ad8a8e7410536566799eea6265f080  skills/engraphis-memory/SKILL.md
 055655db84af07561d002f0c69744313d8413c39f3e873f941f0fa0b1e76dc66  skills/engraphis-memory/references/CONVENTIONS.md
 9d090a03f5b3f36a34d91f66b72c3844591f6915755ac3a6c6ba5f1b16977de5  skills/engraphis-memory/references/SCOPING.md
-4c612f734cc3f50e5566752c6956e43bd432b3c26177e16cde91e196ac9e9b5b  skills/engraphis-memory/references/TOOLS.md
+5edb9bc529339ecff698c1448ce65397a90d7eb4e221864bc265f2bfc329653c  skills/engraphis-memory/references/TOOLS.md

.github/workflows/ci.yml

@@ -52,6 +52,8 @@ jobs:
         run: python -m eval.grounded
       - name: Code-agent arm gate
         run: python -m eval.code_arm
+      - name: Agent specification quality gate
+        run: python -m eval.spec_crawl
       - name: Evidence contract boundary gate
         run: python -m eval.evidence_contracts
 
@@ -129,6 +131,8 @@ jobs:
         run: python -m eval.adversarial_memory_security
       - name: Evidence contract boundary gate
         run: python -m eval.evidence_contracts
+      - name: Agent specification quality gate (numpy-only)
+        run: python -m eval.spec_crawl
       - name: Build and smoke installed core artifacts
         shell: bash
         run: |

AGENTS.md

@@ -59,6 +59,7 @@ python -m eval.reinforcement                                        # bounded re
 python -m eval.adversarial_memory_security                          # prompt/graph boundary
 python -m eval.grounded                                             # grounded-abstain decision gate
 python -m eval.code_arm                                             # coding-agent arm gate
+python -m eval.spec_crawl                                           # agent spec quality & prompt crawl gate
 pyright                                                             # core + backends typecheck
 
 # ── External benchmarks (real numbers need torch + the dataset; see eval/external.py) ──
@@ -294,6 +295,8 @@ worktree is the delivery boundary:
 - **`docs/OBSIDIAN_IMPORT.md`** — the rich Obsidian Markdown adapter (frontmatter, wikilinks,
   aliases, attachments) and its compatibility command. It supplements, rather than replaces,
   the universal document-import guide.
+- **`docs/SPEC_CRAWL.md`** — agent specification quality crawler, 7-axis radar coverage,
+  concept tentacles, "ask, don't guess" clarification flow, and Dashboard Spec Studio.
 - **`AGENTS.md`** (this file) + **`CLAUDE.md`** — how to work in the repo.
 - **`skills/engraphis-memory/`** — portable Agent Skill (SKILL.md + `references/`) that teaches any
   MCP-capable agent the *memory discipline* (when to remember/recall, scoping, tool selection).

BENCHMARKS.md

@@ -94,14 +94,14 @@ interpretation and do not count as additional benchmark-quality gains.
 ### Public numeric evidence registry
 
 Every exact public aggregate retained below comes from the checked-in, public-safe
-[`offline-fixtures-pr253-four-gap-stack-20261008.json`](docs/benchmark-evidence/offline-fixtures-pr253-four-gap-stack-20261008.json) artifact. Its
+[`offline-fixtures-pr260-reviewed-20261011.json`](docs/benchmark-evidence/offline-fixtures-pr260-reviewed-20261011.json) artifact. Its
 SHA-256 is
-`e423e99ba683c1928984903aef715ad97440c9717cecb3bd3f6fe4f6746ea7d0`, also recorded in the
+`b4bf5aa35d84aa7888ebaff61a9c951e849905e5b2f871b5415bdac9e0bad67e`, also recorded in the
 adjacent `.sha256` file. The artifact contains no raw questions, answers, prompts, customer data,
 or per-record content fingerprints.
 
 The fixture-suite digest is
-`e1b7fcea31cb7bf8f8ec85b56e097a0cbfc7d496aef4aef51de73b849b2d17c6`. The artifact defines
+`72ca038b906897a5f1cc97cc6ec0896e79ed4576a8233ef338dc119d23846252`. The artifact defines
 the digest algorithm and records the SHA-256 of every suite and dataset file. Each evidence ID
 also binds its exact command through `sha256(UTF-8 exact command)`:
 
@@ -126,13 +126,13 @@ grounded checks in separate panels; provider billing and MCP transport are not m
 the SVG and matching PNG with:
 
 ```bash
-python scripts/render_benchmark_report.py --report docs/benchmark-evidence/offline-fixtures-pr253-four-gap-stack-20261008.json --output docs/images/context-efficiency.svg --png-output docs/images/context-efficiency.png
+python scripts/render_benchmark_report.py --report docs/benchmark-evidence/offline-fixtures-pr260-reviewed-20261011.json --output docs/images/context-efficiency.svg --png-output docs/images/context-efficiency.png
 ```
 
 The companion examples are also generated from that artifact with:
 
 ```bash
-python -m scripts.render_benchmark_examples --report docs/benchmark-evidence/offline-fixtures-pr253-four-gap-stack-20261008.json --output docs/images/evidence-backed-agent-examples.svg --png-output docs/images/evidence-backed-agent-examples.png
+python -m scripts.render_benchmark_examples --report docs/benchmark-evidence/offline-fixtures-pr260-reviewed-20261011.json --output docs/images/evidence-backed-agent-examples.svg --png-output docs/images/evidence-backed-agent-examples.png
 ```
 The historical-to-executable mapping is in
 [`docs/BENCHMARK_CHANGE_COVERAGE.md`](docs/BENCHMARK_CHANGE_COVERAGE.md).

README.md

@@ -49,7 +49,7 @@ Connect a coding agent over MCP with the [agent setup guide](https://github.com/
 The current registered artifact contains three deterministic offline fixture runs. It separates context size, retrieval quality, and grounded decision checks; these small fixtures do not establish general task performance.
 
 <p align="center">
-  <img src="https://raw.githubusercontent.com/Coding-Dev-Tools/engraphis/522def372c45a26d46bc7a05829964ae0240f3a9/docs/images/context-efficiency.svg" alt="Three registered offline fixtures: structure-aware chunking reduced mean retrieved context from 740.3 to 214.3 tokens per question (71.1%, 18 questions); the serialized JSON-shape proxy fell from 24,590 to 11,138 tokens across 26 payload samples and 260 recalls. Candidate and packed retrieval quality (Recall@5, Hit@5, and answer-token recall) are shown separately (each 1.000). Grounded checks show 5/5 answerable queries grounded and 6/6 abstention queries rejected, including a 1/1 quarantined-evidence probe; 11/11 decisions were correct. MCP transport and provider billing were not measured. Artifact SHA-256 prefix e423e99ba683; see the benchmark guide for the full checksum." width="100%">
+  <img src="https://raw.githubusercontent.com/Coding-Dev-Tools/engraphis/fb992d50da0a14a9ac0cde6e33efdb73c4408c93/docs/images/context-efficiency.svg" alt="Three registered offline fixtures: structure-aware chunking reduced mean retrieved context from 740.3 to 214.3 tokens per question (71.1%, 18 questions); the serialized JSON-shape proxy fell from 24,590 to 11,138 tokens across 26 payload samples and 260 recalls. Candidate and packed retrieval quality (Recall@5, Hit@5, and answer-token recall) are shown separately (each 1.000). Grounded checks show 5/5 answerable queries grounded and 6/6 abstention queries rejected, including a 1/1 quarantined-evidence probe; 11/11 decisions were correct. MCP transport and provider billing were not measured. Artifact SHA-256 prefix b4bf5aa35d84; see the benchmark guide for the full checksum." width="100%">
   <br>
   <sup>Three offline fixtures separate context reduction, candidate and packed retrieval quality, and grounded behavior. The chart shows the artifact checksum prefix; see the benchmark guide for the full checksum and reproduction steps.</sup>
 </p>
@@ -60,7 +60,7 @@ The current registered artifact contains three deterministic offline fixture run
 | Recall payload proxy | JSON-shape proxy: 24,590 → 11,138 tokens (54.71% lower, 26 samples; 260 timed recalls) | Candidate and packed Recall@5, Hit@5, and answer-token recall are each 1.000 |
 | Grounded decisions | 5/5 answerable queries grounded; 6/6 abstention queries rejected, including 1/1 quarantined-evidence check | 11/11 decisions correct |
 
-The payload figure is a serialized JSON-shape estimate, not an MCP transport measurement or provider billing total. See the [Benchmark methodology](https://github.com/Coding-Dev-Tools/engraphis/blob/522def372c45a26d46bc7a05829964ae0240f3a9/BENCHMARKS.md) for artifact identity, counting methods, reproduction commands, external-evaluation boundaries, and limitations.
+The payload figure is a serialized JSON-shape estimate, not an MCP transport measurement or provider billing total. See the [Benchmark methodology](https://github.com/Coding-Dev-Tools/engraphis/blob/fb992d50da0a14a9ac0cde6e33efdb73c4408c93/BENCHMARKS.md) for artifact identity, counting methods, reproduction commands, external-evaluation boundaries, and limitations.
 
 ## Optional Jev assistance
 

docs/ARCHITECTURE_V3.md

@@ -8,7 +8,7 @@ current schema version is 16).
 flowchart LR
     Agent["Agent / host LLM"] --> Intent["remember · link · recall_context (compact) · recall"]
     CLI["engraphis-graph CLI"] --> Service["MemoryService"]
-    MCP["Smart MCP (9 tools) / Classic MCP (39 tools)"] --> Service
+    MCP["Smart MCP (9 tools) / Classic MCP (43 tools)"] --> Service
     HTTP["Dashboard + read-only graph HTTP"] --> Service
     Import["Local resources / PostgreSQL catalog"] --> Extractors["Optional local extractors"]
     Extractors --> Service

docs/KILO_CODE_INTEGRATION.md

@@ -226,7 +226,7 @@ class, and the appropriate executor revalidates all of it before running.
 preserves the named-tool compatibility surface below; new Kilo Code installations should keep the zero-config
 Smart command shown above.
 
-### Classic 39-tool inventory
+### Classic 43-tool inventory
 
 | Category | Tool | What it does |
 |---|---|---|
@@ -269,6 +269,10 @@ Smart command shown above.
 | Ops | `engraphis_check_update` | Check the release source and refresh the persistent update cache. |
 | Maintenance | `engraphis_consolidate` | Pure dry-run or live sweep; structured calls may process a large cluster across retries. |
 | Decision | `engraphis_decide` | Advisory command, contradiction, support, and completion checks. Remote processing requires backend selection and explicit permission for each call. |
+| Quality | `engraphis_spec_crawl` | Multi-dimensional prompt & spec crawler; radar scores, concept tentacles, clarify questions, simulated traces. |
+| Quality | `engraphis_spec_crawl_answer` | Answer a clarification question with user-provided answers, recalculating scores & simulation trace. |
+| Quality | `engraphis_spec_crawl_memories` | Analyze a cluster of memories for contradiction, divergence, and convergence. |
+| Quality | `engraphis_spec_crawl_resolve` | Reconcile divergent or contradicted memories with resolved state and supersede chain. |
 
 ---