feat: complete AMC source review gap receipts

.changeset/amc-a2a-negotiation-methodology.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add A2A-NT-style agent-to-agent negotiation methodology boundaries, migration triggers, and documentation.

.changeset/amc-academiclaw-metric-validity.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add AcademiClaw academic-task metric-validity receipts with fail-closed Score/Shield/Watch surfaces, methodology/docs bindings, and eval-pack proof fields.

.changeset/amc-accessibility-regressions.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Fix accessibility regressions flagged in the Batch 5 audit: top-level CLI `--no-color` handling and help text, accessible names for browser playground and generated console chart canvases, generated-dashboard secondary text contrast, published accessibility statement, and OG image asset verification.

.changeset/amc-accessibility-release-evidence.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add an accessibility release-evidence generator and runbook for recording Playwright axe run status without overclaiming manual assistive-technology coverage.

.changeset/amc-adgen-soc-dataset-replay.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add AD-GEN-style SOC dataset replay receipts, summaries, CI fail-closed fields, and public methodology proof boundaries for ATT&CK-aligned endpoint telemetry benchmark claims.

.changeset/amc-adk-runtime-live-drift.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add ADK TypeScript runtime live-drift receipts so Watch fails closed on missing runtime, framework, graph, tool-registry, eval dataset/case, runner, session, live-queue, API-route, deployment, metric, signed-evidence, and row-hash proof.

.changeset/amc-advanced-rag-notebook-replay.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": minor
+---
+
+Add Advanced RAG notebook replay receipts with course/lesson identity, retrieval variant, notebook/output hashes, environment and dependency-lock hashes, corpus/index/query/reference-answer hashes, retrieval/generation/eval/observability traces, replay command, deterministic seed, query count, RAG triad metric thresholds, CI failed-row reporting, and r42 methodology docs.

.changeset/amc-adversarial-alignment-probes.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add an `adversarialAlignmentProbes` assurance/redteam pack with executable deceptive-alignment, reward-model-gaming, and goal-misgeneralization probes, plus regression coverage and catalog discoverability.

.changeset/amc-agent-belt-methodology-versioning.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add Agent Belt methodology-versioning assurance receipts for reproducible coding-agent evaluation claims.

.changeset/amc-agent-bench-java-coding-validity.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add Agent Bench-style Java coding-agent metric-validity gates so benchmark/source/license, Java task, YAML benchmark, isolated workspace, CLI-agent, cascaded judge, Maven/JUnit/JaCoCo, result, accuracy/pass@k, sample/CI, signed evidence, and row-hash proof fail closed before Java coding-agent benchmark claims are accepted.

.changeset/amc-agent-eval-harness-live-drift.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Adds agent-eval-harness live-drift receipt fields, thresholds, row hashing, Watch alerts, methodology docs, and tests for signed trace/evidence coverage, tool-success, hallucination, latency, cost, framework, trace-mode, and metric-context drift.

.changeset/amc-agent-eval-observability-live-drift.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add fail-closed live-drift proof for agent-evaluation observability rows, including config, telemetry, evidence coverage, metric-set and telemetry distributions, public methodology r133 docs, and source-safe documentation for vladfeigin/llm-agents-evaluation.

.changeset/amc-agent-mont-monitoring-replay.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add Agent_Mont-style monitoring replay receipts to the benchmark corpus, including fail-closed evidence checks for monitoring configuration, framework, token/cost/latency/resource/carbon/log/visualization artifacts, summaries, CI receipt fields, public methodology versioning, and documentation.

.changeset/amc-agent-reading-test-live-drift.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add Agent Reading Test-style web-content reading live-drift receipts with source snapshot, license, homepage, answer key, task manifest, score form, live-site, raw content, canary, alert, signed evidence, and row-hash proof.

.changeset/amc-agent-security-live-drift.md

@@ -0,0 +1,7 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add agent-security control live-drift receipts for Watch and Shield.
+
+Live drift rows can now bind guard identity, policy hashes, taint/proxy/audit/telemetry/eval-pack/classifier proof, origin and taint coverage, policy-decision accuracy, secret-scrub rate, audit integrity, attack-effectiveness rate, false-positive rate, guard latency, signed evidence, and row hashes. Missing or degraded agent-security evidence fails closed through score drift, behavior drift, Watch alerts, Shield verification, and public methodology r68.

.changeset/amc-agent-testing-methodology-live-drift.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add agent-testing methodology live-drift receipts for Watch and Shield. Live rows can now carry testing taxonomy, methodology, scenario, fault-injection, observability, safety, standards, category, approach, fault-model, benchmark-family, coverage, resilience, safety-regression, and observability-signal evidence so AMC fails closed when live traffic drifts away from the declared testing methodology despite stable generic scores.

.changeset/amc-agent-workflow-kit-replay.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add Agent Workflow Kit-style workflow replay receipts with fail-closed source, policy, approval, verification, docs-check, replay, threshold, CI receipt, methodology, and documentation coverage.

.changeset/amc-agentbench-config-replay.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Adds AgentBench-style config-pinned replay receipts to benchmark corpus runs so source, repository, dataset, agent/global/model-server/environment/dependency, run/replay command, trace, result, metric, seed, sample, shuffle, replay-pass, trace-coverage, signed-evidence, and row-hash proof fail closed.

.changeset/amc-agentdefense-provider-drift.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add AgentDefense-Bench provider-drift receipts with source/MCP/security-defense proof, fail-closed Watch/CI alerts, public methodology r200 binding, and docs for the live verified source boundary.

.changeset/amc-agentest-scenario-metric-validity.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": minor
+---
+
+Add Agentest-style scenario-test metric-validity proof with signed source, endpoint, scenario, persona, goal, knowledge, tool-mock, scripted-turn, trajectory assertion, LLM judge, comparison, CI reporter, result, sample-size, confidence-interval, and row-hash evidence. Missing or invalid proof now fails closed when `requireAgentScenarioTestProof` is enabled.

.changeset/amc-agentic-graph-rag-metric-validity.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add Agentic Graph RAG metric-validity receipts with fail-closed source/no-license, graph/RAG, vector-store, evaluation, experiment-tracking, UI-question, dependency-lock, owner, confidence-interval, signed-evidence, artifact-hash, and row-hash proof.

.changeset/amc-agentic-search-live-drift.md

@@ -0,0 +1,7 @@
+---
+"agent-maturity-compass": minor
+---
+
+Add agentic-search live drift receipts for baseline-to-live monitoring.
+
+Rows can now bind benchmark id, dataset family, query type, query/task ids, source and tool-config hashes, planner/search/citation/synthesis trace hashes, result manifest hash, planning/query-decomposition/relevance/synthesis scores, and citation coverage. Score, Watch, and Shield now fail closed on agentic-search score drops, citation or trace coverage gaps, dataset-family drift, query-type drift, tool-context drift, missing signed evidence, or receipt hash mismatches.

.changeset/amc-agentkernelarena-gpu-kernel-replay.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add AgentKernelArena-style GPU-kernel replay receipts with task/config, agent roster, workspace isolation, GPU profile, compile/correctness/performance proof, speedup delta, replay/result coverage, CI receipt, signed evidence, and row-hash gates.

.changeset/amc-agentrial-statistical-question-explainability.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add AgentTrial-style statistical question-explainability receipts with repeated trial counts, Wilson confidence intervals, bootstrap cost/latency, failure attribution, regression comparison, CI proof, reliability score, and row-hash gates.

.changeset/amc-agentrim-diagnostic-question.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add an AgenTRIM-backed diagnostic question for per-step least-privilege tool access with status-aware validation evidence gates.

.changeset/amc-agentstock-judge-calibration.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add AgentStock-style future-outcome ranking proof to judge calibration receipts, including source snapshot, leaderboard, PnL, appeal, replay, signed-evidence, and Watch fail-closed gates.

.changeset/amc-agiflow-observability-methodology.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": minor
+---
+
+Add LLM workflow observability methodology-versioning boundaries for trace, visual-debugger, prompt/model registry, frontend analytics, user-feedback, session-replay, telemetry privacy, migration, badge, and report proof.

.changeset/amc-ai-agent-benchmark-comparison-replay.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add AI-agent benchmark comparison replay proof to the benchmark corpus receipt so source, repository, license, agent roster, benchmark dataset, source/pricing/user-report/leaderboard/score manifests, eval-pack, fixture, replay, result, score-delta report, CI receipt, coverage metrics, signed evidence, and row hashes fail closed before comparison claims are accepted.

.changeset/amc-ai-coding-landscape-explainability.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add AI-coding landscape question-explainability lenses. Question receipts can now bind source category, dataset refs and SHA-256 hashes, update cadence, freshness, cohort refs, benchmark/tool/model refs, accepted evidence, rejected-evidence reasons, and repair hints into row hashes, fail-closed status, Studio evidence drilldown, and public methodology r36.

.changeset/amc-ai-evaluation-guide-methodology.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add Awesome AI Evaluation Guide public-methodology receipts with source/license, default branch, guide manifests, benchmark/tool taxonomies, metric-selection, threshold, calibration, trace, human-review, cost-control, deprecation, migration, signed evidence, and row-hash proof.

.changeset/amc-ai-reputation-claude-live-drift.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add AI Reputation Claude live-drift receipts with source/no-license, agent roster, skill catalog, review-source, sentiment, competitor, response-policy, crisis, report, baseline/live result, drift statistic, alert receipt, brand-safety metric, signed-evidence, and row-hash proof.

.changeset/amc-aicrypto-methodology-boundary.md

@@ -0,0 +1,7 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add an AICrypto-style cryptography benchmark methodology boundary.
+
+Public cryptography capability claims now require methodology-versioned proof for paper/dataset versions, MCQ/CTF/proof task families, expert baselines, sandbox/toolchain evidence, proof rubrics, scoring formulas, thresholds, signed evidence, and row hashes before AMC reports or badges can use those claims as external evidence.

.changeset/amc-alignment-feedback-source-validation.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add alignment feedback-source validation scoring and a diagnostic question for evaluator source quality, bias, collusion, and signed feedback provenance.

.changeset/amc-alignment-index-subcategories.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Expose alignment-index subcategory breakdowns for goal misgeneralization, reward hacking, deceptive alignment, feedback source validation, sycophancy, and sabotage.

.changeset/amc-api-key-cli.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add `amc api key create`, `amc api key list`, and `amc api key revoke` for local programmatic API key management. Keys are displayed only once on creation; the persisted store keeps hashed secret material and public metadata under `.amc/auth/api-keys.json`.

.changeset/amc-api-operational-contracts.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Document API rate-limit headers, org SSE reconnect behavior, and webhook retry boundaries; add org SSE event IDs and a 15-second reconnect hint.

.changeset/amc-assurance-certificate-threshold-guide.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Document signed assurance certificate issuance, verification, and policy threshold tuning.

.changeset/amc-assurance-demo-nosign.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add `amc assurance run --demo --no-sign` and make single-pack no-sign runs vault-less.

.changeset/amc-assurance-lab-index-mjs-docs.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add regression coverage for Assurance Lab pack-authoring docs so community packs document `index.mjs` as the scaffolded entry point and `index.js` only as legacy fallback.

.changeset/amc-assurance-remediation-priority.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add a remediation-priority section to text assurance runs so failed scenarios are ordered by severity with reason, fix hint, evidence path, and verbose rerun command.

.changeset/amc-assurance-run-nosign-audit.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add built-CLI coverage for `assurance run --all --no-sign` and align the UX audit with the current unsigned assurance behavior.

.changeset/amc-awesome-agent-memory-live-drift.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add Awesome-Agent-Memory-style memory-catalog live-drift receipts with source snapshot, no-license boundary, README blob, taxonomy, benchmark/eval, drift statistic, alert, signed evidence, and row-hash proof.

.changeset/amc-azure-agent-lab-replay-corpus.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add Azure Agent Lab replay-corpus receipts with lab/module identity, workshop and notebook hashes, Azure service/project/search/RAG/tool/evaluator configs, cloud-run and identity proof, replay command hashes, deterministic seeds, scenario counts, evaluation scores, groundedness thresholds, CI failed-row ids, Watch alerts, and public methodology/docs coverage.

.changeset/amc-backdooragent-stage-backdoor-live-drift.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add BackdoorAgent-style stage-aware backdoor live-drift receipts so attack success, clean accuracy, trigger persistence, trigger propagation, trajectory coverage, evidence coverage, stage/task/attack-family drift, signed evidence, and row hashes fail closed.

.changeset/amc-batch5-audit-score-consistency.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add a Batch 5 audit consistency regression so persona table scores, section headings, and rating lines stay aligned after follow-up fixes.

.changeset/amc-benchloop-local-benchmark-replay.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add BenchLoop-style local benchmark replay receipts with fail-closed suite, harness, provider, hardware, run, trace, latency, token, export, and metric evidence bindings.

.changeset/amc-benchmark-hackability-audit-replay.md

@@ -0,0 +1,7 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add benchmark-hackability audit replay receipts for replay benchmark corpora.
+
+Replay rows can now bind scanner identity, target benchmark/task manifests, audit configs, phase traces, static-tool reports, AI-inspection traces, vulnerability finding manifests, dashboard/report artifacts, replay commands, sandbox controls, PoC validation, vulnerability-class coverage, task-count coverage, exploitability thresholds, signed evidence, and row hashes. Missing benchmark-hackability evidence fails closed through the manifest, CI receipt, Shield verification, Watch alerts, and public methodology r67.

.changeset/amc-benchmark-submission-question-explainability.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add benchmark-submission question explainability receipts with task status, criterion scoring, leaderboard metric views, replay hashes, fail-closed validation, drilldown previews, guide remediation hints, and public methodology r65.

.changeset/amc-besttester-replay-corpus.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add BestTester replay-corpus receipts for QA-agent benchmark evidence. AMC now validates source/license snapshots, package and lockfile refs, Playwright and TypeScript proof, test/agent/MCP/security/workflow artifacts, LLM-judge agreement, security coverage, CI coverage, signed evidence, and row hashes before BestTester-style claims can pass Score, Shield, or Watch gates.

.changeset/amc-bioagentbench-metric-validity.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add BioAgentBench-style bioinformatics agent metric-validity proof for Score/Shield rows. AMC now requires signed benchmark/source, task, input dataset, truth/reference, workflow reproduction, Docker/environment, tool-version, harness, grader, result artifact, perturbation, privacy-boundary, owner, sample-size, confidence-interval, and row-hash evidence before bioinformatics workflow claims can be used externally.

.changeset/amc-biokgbench-biomedical-kg-replay.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add BioKGBench-style biomedical KG replay receipts so source, repository, paper, license, dataset release, knowledge graph, KGCheck/KGQA/SCV task manifests, agent/RAG/Neo4j configs, evaluation scripts, result manifests, error-discovery reports, replay commands, CI receipts, deterministic seeds, metrics, thresholds, signed evidence, and row hashes fail closed before public biomedical KG benchmark claims.

.changeset/amc-biomedarena-biomedical-replay.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add BioMedArena-style biomedical harness replay receipts with source, harness, benchmark-family, tool-mode, adapter/tool/vendor, baseline, result, replay, CI, coverage, sandbox, and row-hash proof.

.changeset/amc-board-l3-risk-memo.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add a board-facing L3 business-risk memo and link it from executive surfaces without over-approving production use.

.changeset/amc-business-fair-scenario.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add `amc business fair-scenario` to run a deterministic FAIR-style scenario loss distribution with explicit frequency and loss-magnitude calibration ranges, maturity-adjusted exposure, P10/P50/P90/P95 outputs, and risk-appetite status without claiming certified Open FAIR or native GRC sync.

.changeset/amc-business-grc-export.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add `amc business grc-export` to turn portfolio maturity-risk inputs into CSV, JSON, or Markdown GRC treatment-plan exports with owner, due-date, risk appetite, ISO 31000 context, and FAIR-style loss-frequency/loss-magnitude fields.

.changeset/amc-business-risk-heatmap.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add `amc business heatmap` for portfolio-level monetary risk heatmaps across agents, business units, residual expected annual loss, and risk-appetite breaches.

.changeset/amc-business-risk-quantification.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add `amc business risk` to estimate residual incident frequency, expected annual loss, expected loss reduction, and risk-appetite status from agent maturity.

.changeset/amc-business-roi-calculator.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add `amc business roi` for cost-of-trust-gap ROI estimates backed by the maturity-linked expected annual loss model.

.changeset/amc-buyer-packages-pricing-link.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Link buyer packages prominently from the homepage pricing section for procurement workflows.

.changeset/amc-calibra-public-methodology.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add Calibra-style public-methodology receipts with campaign matrix, task, report, dashboard, changelog, migration, and row-hash proof.

.changeset/amc-catastrophic-risk-indicators.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add ForesightSafety catastrophic-risk scoring for self-replication, resource acquisition, shutdown resistance, persistence, goal-preservation pressure, and cross-system propagation.

.changeset/amc-cc-plugin-eval-metric-validity.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add cc-plugin-eval-style metric-validity proof fields, fail-closed thresholds, public methodology boundaries, and documentation for component-trigger reliability evidence.

.changeset/amc-cert-preview-no-sign.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add `amc cert generate --no-sign` / `--preview` for unsigned trust-certificate previews that work without vault signing, are labeled `UNSIGNED_PREVIEW`, and are intentionally rejected by verifier logic until regenerated as signed certificates.

.changeset/amc-chaos-reliability-live-drift.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add chaos-reliability live-drift receipts for Watch and Shield. Live rows can now carry benchmark, scenario, chaos profile, injection, mutation, endpoint contract, judge, trace bundle, score ledger, agent-card, improvement-eval, framework, modality, benchmark-family, production-reliability, resilience, chaos-drop, recovery, and failure-trace evidence so AMC fails closed when live agents drift under failure-injection pressure despite stable generic scores.

.changeset/amc-chipbenchmark-metric-validity.md

@@ -0,0 +1,5 @@
+---
+"agent-maturity-compass": patch
+---
+
+Add ChipBenchmark-style hardware benchmark metric-validity receipts across Score, Shield, Watch, public methodology, and docs. AMC now fails closed unless ChipBenchmark claims bind source snapshot, no-license boundary, benchmark/hardware/model/precision manifests, environment and runner/serving scripts, result/frontend/pricing datasets, throughput/latency/cost metrics, regression thresholds, owners, confidence intervals, signed evidence, and row hashes.