cool-workflow 0.1.78

package/docs/multi-agent-cli-mcp-surface.7.md

@@ -0,0 +1,265 @@
+# Multi-Agent CLI + MCP Surface
+
+CW v0.1.20 adds the preferred host-facing control loop for multi-agent work:
+
+```text
+multi-agent run -> status -> step -> blackboard -> score -> select
+```
+
+CW v0.1.25 extends this surface with State Explosion Management commands:
+`summary refresh`, `summary show`, `blackboard summarize`,
+`multi-agent summarize`, and `multi-agent graph --view <view>` (with optional
+`--focus <id>` and `--depth <n>`). Matching MCP tools are `cw_summary_refresh`,
+`cw_summary_show`, `cw_blackboard_summarize`, `cw_multi_agent_summarize`, and
+`cw_multi_agent_graph_compact`. All responses keep source refs and expansion
+hints. See [state-explosion-management.7.md](state-explosion-management.7.md).
+
+CW v0.1.26 adds `multi-agent reasoning <run-id> [--evidence <id>] [--refresh]`
+(MCP: `cw_evidence_reasoning`, `cw_evidence_reasoning_refresh`), which explains
+*why* each evidence item was adopted, and an additive `rationaleStatus` field on
+`multi-agent evidence` rows. See
+[evidence-adoption-reasoning-chain.7.md](evidence-adoption-reasoning-chain.7.md).
+
+This is userland over the existing kernel records. The low-level topology,
+multi-agent, blackboard, candidate, audit, and commit primitives remain
+available, but agent hosts should use this high-level surface when driving a
+run.
+
+## CLI Loop
+
+Create or attach a topology-backed run without spawning workers:
+
+```bash
+node scripts/cw.js multi-agent run <run-id> --topology judge-panel --task <task-id>
+node scripts/cw.js multi-agent run --app architecture-review --repo /path/to/repo --question "Review this" --topology map-reduce
+```
+
+Read the combined host status:
+
+```bash
+node scripts/cw.js multi-agent status <run-id>
+node scripts/cw.js multi-agent status <run-id> --json
+```
+
+Perform one deterministic step at a time:
+
+```bash
+node scripts/cw.js multi-agent step <run-id> --sandbox readonly
+```
+
+`step` may create a dispatch manifest, collect fanin, snapshot the blackboard,
+register a candidate, score a candidate with existing verifier evidence, select
+a scored candidate, or recommend the verifier-gated commit command. It never
+spawns agents directly.
+
+Work with the active blackboard when it is unambiguous:
+
+```bash
+node scripts/cw.js multi-agent blackboard <run-id> summary
+node scripts/cw.js multi-agent blackboard <run-id> topics
+node scripts/cw.js multi-agent blackboard <run-id> post --topic <topic-id> --body "finding" --evidence <ref>
+node scripts/cw.js multi-agent blackboard <run-id> add-artifact --topic <topic-id> --kind worker-result --path result.md
+node scripts/cw.js multi-agent blackboard <run-id> snapshot
+```
+
+Score and select explicitly:
+
+```bash
+node scripts/cw.js multi-agent score <run-id> <candidate-id> --criterion correctness=1 --criterion evidence=1 --evidence <ref>
+node scripts/cw.js multi-agent select <run-id> <candidate-id> --score <score-id> --reason "verifier-backed candidate"
+node scripts/cw.js commit <run-id> --selection <selection-id> --reason "verified winner"
+```
+
+## Operator Inspection
+
+v0.1.21 extends the host loop with focused operator commands:
+
+```bash
+node scripts/cw.js multi-agent graph <run-id>
+node scripts/cw.js multi-agent dependencies <run-id>
+node scripts/cw.js multi-agent failures <run-id>
+node scripts/cw.js multi-agent evidence <run-id>
+```
+
+The human output is compact and operational: agent graph, dependencies, failed
+or blocked agents, adopted evidence, missing evidence, and the next action.
+Use `--json` or `--format json` for deterministic script output.
+
+## MCP Tools
+
+MCP hosts should prefer:
+
+- `cw_multi_agent_run`
+- `cw_multi_agent_status`
+- `cw_multi_agent_step`
+- `cw_multi_agent_blackboard`
+- `cw_multi_agent_score`
+- `cw_multi_agent_select`
+- `cw_multi_agent_graph`
+- `cw_multi_agent_dependencies`
+- `cw_multi_agent_failures`
+- `cw_multi_agent_evidence`
+
+The older `cw_multi_agent_*`, `cw_topology_*`, `cw_blackboard_*`, and
+`cw_candidate_*` tools remain advanced primitives.
+
+## Stable Responses
+
+Every high-level response is JSON and includes:
+
+- `runId`
+- active topology and multi-agent ids
+- blackboard and topic ids
+- candidate, selection, commit, and audit ids
+- `state`, `performed`, `nextAction`, and `nextActions`
+- `blockedReasons`, `requiredHostAction`, and `evidenceRequirements`
+- state, report, blackboard, audit, ranking, worker manifest, and result paths
+- combined topology, multi-agent, multi-agent operator, blackboard, worker,
+  candidate, feedback, commit, and audit summaries
+
+## Fail-Closed Rules
+
+The host surface fails closed when:
+
+- active topology or blackboard state is ambiguous
+- a fanout has incomplete role coverage
+- worker output has not been recorded
+- fanin lacks required evidence or blackboard links
+- score evidence is missing
+- selection lacks score or verifier readiness
+- a verifier-gated commit is not ready
+
+## Smoke Coverage
+
+`test/multi-agent-cli-mcp-surface-smoke.js` covers the full host loop over the
+official `judge-panel` topology, CLI and MCP parity, ambiguous topology
+failure, missing evidence failure, successful score/select, blackboard
+artifact/message linkage, audit provenance, and Operator UX next actions. It is
+included in `npm test` and `npm run release:check`.
+
+`test/multi-agent-operator-ux-smoke.js` covers the v0.1.21 graph,
+dependencies, failures, evidence adoption, report output, and MCP parity.
+
+`test/multi-agent-trust-policy-audit-smoke.js` covers the v0.1.22
+role-policy, blackboard-write, message-provenance, judge-rationale,
+policy-violation, report, audit provenance, and MCP parity surface.
+
+`test/multi-agent-eval-replay-harness-smoke.js` covers the v0.1.24 eval/replay
+commands and MCP tools: snapshot, replay, compare, score, gate, report, and
+controlled regression detection.
+## CLI ↔ MCP Parity (v0.1.27)
+
+Every command and tool referenced above is declared in the v0.1.27 capability
+registry (`src/capability-registry.ts`) and validated by `npm run parity:check`,
+so `cw <cmd> --json` and the matching `cw_<tool>` result render one data source.
+See [cli-mcp-parity.7.md](cli-mcp-parity.7.md).
+
+## Run Registry / Control Plane (v0.1.28)
+
+The runs described here are indexed, searchable, resumable, archivable, and
+rerunnable across repos by the v0.1.28 Run Registry / Control Plane, which derives
+a fingerprinted, fail-closed index over the same per-run `.cw/runs/<id>/state.json`
+source of truth. See [run-registry-control-plane.7.md](run-registry-control-plane.7.md).
+
+## Execution Backends (v0.1.29)
+
+v0.1.29 lifts execution into a pluggable driver layer: one narrow `ExecutionBackend`
+contract with interchangeable `node`/`bun`/`shell`/`container`/`remote`/`ci`
+drivers, selected by `--backend` (parallel to `--sandbox`) and inspected via
+`backend list|show|probe`. The result/evidence envelope is schema-identical across
+backends; the backend id + sandbox attestation are recorded as provenance, so this
+surface is unchanged regardless of which backend executed a run. See
+[execution-backends.7.md](execution-backends.7.md).
+## Web / Desktop Workbench (v0.1.30)
+
+v0.1.30 adds the Web / Desktop Workbench: a read-only, localhost-only human
+console that renders this surface (and the other four operator panels — run
+graph, blackboard, worker logs, candidate compare, audit timeline) for any run,
+reading the SAME capability `--json` payloads. It is a THIRD FRONT DOOR alongside
+the CLI and MCP that holds no authoritative state and forks no schema: each panel
+equals its `cw <cmd> --json` payload byte-for-byte (parity-gated), and refresh
+re-derives everything from disk. See
+[web-desktop-workbench.7.md](web-desktop-workbench.7.md).
+
+## Observability + Cost Accounting (v0.1.31)
+
+v0.1.31 adds Observability + Cost Accounting: `metrics show`/`metrics summary`
+derive durations, failure/verifier/acceptance rates (with sample counts and
+fail-closed `n/a`), and host-attested token/cost from existing durable run state
+— no metrics database, no collector daemon, no hidden counter. Usage is additive
+and optional (absent ⇒ `unreported`, never 0); cost is `attested` (attested usage
+× a recorded pricing policy) or clearly `estimated`, with pricing as policy. Both
+verbs are parity-gated and render read-only in the v0.1.30 Workbench. See
+[observability-cost-accounting.7.md](observability-cost-accounting.7.md).
+
+
+## Team Collaboration (v0.1.32)
+
+v0.1.32 adds Team Collaboration: a host-attested actor and append-only
+approvals/rejections/comments/handoffs provenance-linked to a durable target,
+plus a review gate that STACKS ON the verifier gate — required approvals from
+authorized roles, enforced inside `resolveCommitGate` AFTER the verifier checks
+and never instead of them, failing closed on quorum/authority/self-approval and
+recording who approved the very artifact that shipped. Policy (required approvals,
+authorized roles, self-approval) is data, default off (pre-v0.1.32 behavior
+unchanged). The verbs are parity-gated and render read-only in the v0.1.30
+Workbench. See [Team Collaboration](team-collaboration.7.md).
+
+## Release Tooling (v0.1.33)
+
+the per-tag mechanical surfaces (version bump across 17 surfaces, feature scaffold, and the forward-reference docs) become deterministic scripts, with a de-duplicated release gate. See release-tooling(7).
+
+## Real Execution Backend Integrations (v0.1.34)
+
+container/remote/ci backends really execute (docker/podman run, remote/CI POST-and-poll) under the sandbox contract, with byte-stable evidence vs node and fail-closed refusal when a runtime/endpoint is unavailable. See real-execution-backends(7).
+
+## Node Snapshot / Diff / Replay (v0.1.35)
+
+per-node snapshot, structural diff, and isolated deterministic replay over StateNode, reusing the v0.1.23 eval harness; fail-closed on source drift (valid|stale|absent). See node-snapshot-diff-replay(7).
+
+## Contract Migration Tooling (v0.1.36)
+
+first-class declared migration registry (run-state + workflow-app) with per-edge compatibility proofs, fail-closed reachability, and a round-trip/non-destruction prover. See contract-migration-tooling(7).
+
+## Control-Plane Scheduling (v0.1.37)
+
+priority + concurrency limits + lease lifecycle + retry/backoff + fail-closed park over the v0.1.28 Run Registry queue; policy-as-data, deterministic. See control-plane-scheduling(7).
+
+## Agent Delegation Drive (v0.1.38)
+
+spawn an external agent process per worker, capture result.md + attestation, auto-drive plan->dispatch->fulfill->accept->commit
+
+## Run Retention & Provable Reclamation (v0.1.39)
+
+tiered, append-only, cryptographically-verifiable run reclamation: seal the audit skeleton, free the reconstructable bulk, prove it
+
+## Durable State & Locking (v0.1.40)
+
+atomic temp->rename writes + fsync-durability for authoritative stores; portable stale-stealing file lock serializing the cross-process read-modify-write stores
+
+## Self-Audit Hardening & Pure-Router Decomposition (v0.1.41)
+
+evidence grounding + durable audit append + symlink-hardened containment + deterministic worker ids + recursive redaction; BackendRegistry self-describing drivers (no per-id switches); orchestrator god-object decomposed into per-domain operation modules (pure loadRun->delegate router)
+
+## Robust Result Ingest (v0.1.42)
+
+capture findings/evidence from any reasonable agent shape (alt keys + prose), CW derives grounded evidence itself, warn on empty capture — closes the v0.1.41 live-drive 'accepted with 0 captured' failure
+
+## No-False-Green Gate & Launch Prep (v0.1.43)
+
+Hard gate blocking empty-capture verifier-gated commits, plus quickstart and launch-prep docs.
+
+## Release-Gate Determinism & Agents Vendor (v0.1.44)
+
+Release-readiness checks now validate the committed blob (`git show HEAD:<path>`) instead of the mutable working tree — eliminating false-red/false-green from concurrent working-tree writes (iCloud/Spotlight/editor). Adds the `agents` vendor manifest target: a generated `.agents/plugins/cool-workflow/` adapter giving any non-Claude AI agent one common interface to CW.
+
+## P1-P2 Fixes & CI Content Surfaces (v0.1.49)
+
+Migration DAG with reversible edges (v0.1.45), capability auto-discovery (v0.1.46), vendor-adapter registry (v0.1.47), state auto-compaction and P2 fixes (v0.1.48), plus CI content-surface determinism hardening (v0.1.49).
+0.1.51
+
+0.1.76
+
+0.1.77
+
+0.1.78

package/docs/multi-agent-eval-replay-harness.7.md

@@ -0,0 +1,302 @@
+# Multi-Agent Eval & Replay Harness
+
+CW v0.1.23 added a deterministic replay harness for topology-backed
+multi-agent runs. It turns a completed run into plain JSON evidence that can be
+replayed without live agents, compared with normalized rules, scored, and used
+as a release gate.
+
+CW v0.1.25 extends the harness with State Explosion Management metrics so the
+derived summary layer is regression-gated alongside the raw run:
+`summary_freshness`, `compact_graph_parity`, `blackboard_digest_parity`,
+`critical_path_parity`, `evidence_digest_parity`, and `expansion_ref_integrity`.
+Pre-0.1.25 snapshots load with empty summary sections, so old fixtures stay
+backward compatible. See
+[state-explosion-management.7.md](state-explosion-management.7.md).
+
+CW v0.1.26 adds Evidence Adoption Reasoning Chain metrics: `reasoning_freshness`,
+`reasoning_chain_parity`, and `reasoning_unexplained_parity`. Pre-0.1.26
+snapshots load with empty reasoning sections. See
+[evidence-adoption-reasoning-chain.7.md](evidence-adoption-reasoning-chain.7.md).
+
+The harness is intentionally file-first:
+
+- snapshots, replay runs, comparisons, scores, findings, gates, and reports are
+  stored under `.cw/evals/<suite-id>/`
+- the baseline run is not mutated during replay
+- replay output is written to an isolated `replay/` directory
+- every CLI command supports deterministic JSON with `--json` or
+  `--format json`
+- MCP tools return JSON only and include generated artifact paths
+
+## Commands
+
+Create a snapshot from a multi-agent run:
+
+```bash
+node scripts/cw.js eval snapshot <run-id> --id <suite-id>
+node scripts/cw.js eval snapshot <run-id> --id <suite-id> --json
+```
+
+Replay without live agents:
+
+```bash
+node scripts/cw.js eval replay .cw/evals/<suite-id>/snapshot.json
+```
+
+Compare, score, gate, and report:
+
+```bash
+node scripts/cw.js eval compare \
+  .cw/evals/<suite-id>/snapshot.json \
+  .cw/evals/<suite-id>/replay-run.json
+
+node scripts/cw.js eval score .cw/evals/<suite-id>/replay-run.json
+node scripts/cw.js eval gate .cw/evals/<suite-id>
+node scripts/cw.js eval report .cw/evals/<suite-id>/replay-run.json
+```
+
+`npm run eval:replay` runs the deterministic smoke suite and is included in
+`npm test` and `npm run release:check`.
+
+Human output uses stable panels:
+
+```text
+Eval Suite
+Replay Status
+Graph Comparison
+Evidence Comparison
+Trust / Policy / Audit Comparison
+Candidate Score Comparison
+Selection / Commit Gate
+Regression Findings
+Final Verdict
+Next Action
+```
+
+## Artifacts
+
+Each suite writes predictable files:
+
+- `suite.json`
+- `snapshot.json`
+- `replay-run.json`
+- `comparison.json`
+- `score.json`
+- `findings.json`
+- `gate.json`
+- `report.md`
+
+The snapshot captures workflow app identity, inputs, topology shape, roles,
+groups, memberships, fanout/fanin state, blackboard records, worker outputs,
+candidate scores, selection rationale, verifier-gated commit inputs,
+trust/policy/audit records, expected operator summaries, evidence adoption, and
+report sections.
+
+## Comparison Rules
+
+The comparison checks:
+
+- topology id and topology run shape
+- roles, groups, memberships, fanout, and fanin records
+- dependency edges and failure rows
+- blackboard records and message provenance
+- role policies, permission decisions, write audit, judge rationale, panel
+  decisions, and policy violations
+- evidence adoption status
+- candidate scores, selected candidate, and selection rationale
+- verifier-gated commit readiness
+- report sections
+
+Normalization removes unstable paths, timestamps, generated temp roots, and
+machine-local directories. It does not hide changed evidence, policy,
+selection, scoring, or commit-gate behavior.
+
+## Scoring
+
+Scores are deterministic metrics:
+
+- `replay_completed`
+- `graph_parity`
+- `role_parity`
+- `group_parity`
+- `membership_parity`
+- `fanout_parity`
+- `fanin_parity`
+- `dependency_parity`
+- `failure_parity`
+- `blackboard_record_parity`
+- `evidence_adoption_parity`
+- `trust_audit_parity`
+- `role_policy_parity`
+- `permission_decision_parity`
+- `policy_violation_parity`
+- `blackboard_provenance_parity`
+- `judge_rationale_parity`
+- `panel_decision_parity`
+- `candidate_score_parity`
+- `selection_parity`
+- `verifier_commit_gate_parity`
+- `report_parity`
+
+Each metric returns `id`, `status`, `score`, `maxScore`, `reason`, evidence
+refs, baseline refs, and replay refs.
+
+## Gate
+
+`eval gate` fails closed when replay artifacts are missing or when comparison
+findings show a regression. This includes missing judge rationale, changed
+selected candidate, changed evidence adoption, changed policy violations,
+missing provenance, lost verifier-gated commit readiness, or graph/dependency
+loss.
+
+Improvements can be represented as changed findings in future suites, but they
+must be visible in `score.json`, `findings.json`, and `report.md` before a
+release gate can accept them.
+
+## MCP Parity
+
+The MCP surface mirrors the CLI:
+
+- `cw_eval_snapshot`
+- `cw_eval_replay`
+- `cw_eval_compare`
+- `cw_eval_score`
+- `cw_eval_gate`
+- `cw_eval_report`
+
+MCP responses are deterministic JSON and include artifact paths.
+
+## Release Use
+
+Use this harness after a topology-backed run reaches score, selection, and a
+verifier-gated commit:
+
+```bash
+node scripts/cw.js eval snapshot <run-id> --id release-replay
+node scripts/cw.js eval replay .cw/evals/release-replay/snapshot.json
+node scripts/cw.js eval compare .cw/evals/release-replay/snapshot.json .cw/evals/release-replay/replay-run.json
+node scripts/cw.js eval score .cw/evals/release-replay/replay-run.json
+node scripts/cw.js eval gate .cw/evals/release-replay
+node scripts/cw.js eval report .cw/evals/release-replay/replay-run.json
+```
+
+The gate proves the replay completed, graph/dependencies stayed stable,
+evidence adoption stayed traceable, trust/policy/audit records remained
+explainable, judge rationale is present, scoring/selection did not regress, and
+verifier-gated commit readiness still holds.
+## CLI ↔ MCP Parity (v0.1.27)
+
+Every command and tool referenced above is declared in the v0.1.27 capability
+registry (`src/capability-registry.ts`) and validated by `npm run parity:check`,
+so `cw <cmd> --json` and the matching `cw_<tool>` result render one data source.
+See [cli-mcp-parity.7.md](cli-mcp-parity.7.md).
+
+## Run Registry / Control Plane (v0.1.28)
+
+The runs described here are indexed, searchable, resumable, archivable, and
+rerunnable across repos by the v0.1.28 Run Registry / Control Plane, which derives
+a fingerprinted, fail-closed index over the same per-run `.cw/runs/<id>/state.json`
+source of truth. See [run-registry-control-plane.7.md](run-registry-control-plane.7.md).
+
+## Execution Backends (v0.1.29)
+
+v0.1.29 lifts execution into a pluggable driver layer: one narrow `ExecutionBackend`
+contract with interchangeable `node`/`bun`/`shell`/`container`/`remote`/`ci`
+drivers, selected by `--backend` (parallel to `--sandbox`) and inspected via
+`backend list|show|probe`. The result/evidence envelope is schema-identical across
+backends; the backend id + sandbox attestation are recorded as provenance, so this
+surface is unchanged regardless of which backend executed a run. See
+[execution-backends.7.md](execution-backends.7.md).
+## Web / Desktop Workbench (v0.1.30)
+
+v0.1.30 adds the Web / Desktop Workbench: a read-only, localhost-only human
+console that renders this surface (and the other four operator panels — run
+graph, blackboard, worker logs, candidate compare, audit timeline) for any run,
+reading the SAME capability `--json` payloads. It is a THIRD FRONT DOOR alongside
+the CLI and MCP that holds no authoritative state and forks no schema: each panel
+equals its `cw <cmd> --json` payload byte-for-byte (parity-gated), and refresh
+re-derives everything from disk. See
+[web-desktop-workbench.7.md](web-desktop-workbench.7.md).
+
+## Observability + Cost Accounting (v0.1.31)
+
+v0.1.31 adds Observability + Cost Accounting: `metrics show`/`metrics summary`
+derive durations, failure/verifier/acceptance rates (with sample counts and
+fail-closed `n/a`), and host-attested token/cost from existing durable run state
+— no metrics database, no collector daemon, no hidden counter. Usage is additive
+and optional (absent ⇒ `unreported`, never 0); cost is `attested` (attested usage
+× a recorded pricing policy) or clearly `estimated`, with pricing as policy. Both
+verbs are parity-gated and render read-only in the v0.1.30 Workbench. See
+[observability-cost-accounting.7.md](observability-cost-accounting.7.md).
+
+
+## Team Collaboration (v0.1.32)
+
+v0.1.32 adds Team Collaboration: a host-attested actor and append-only
+approvals/rejections/comments/handoffs provenance-linked to a durable target,
+plus a review gate that STACKS ON the verifier gate — required approvals from
+authorized roles, enforced inside `resolveCommitGate` AFTER the verifier checks
+and never instead of them, failing closed on quorum/authority/self-approval and
+recording who approved the very artifact that shipped. Policy (required approvals,
+authorized roles, self-approval) is data, default off (pre-v0.1.32 behavior
+unchanged). The verbs are parity-gated and render read-only in the v0.1.30
+Workbench. See [Team Collaboration](team-collaboration.7.md).
+
+## Release Tooling (v0.1.33)
+
+the per-tag mechanical surfaces (version bump across 17 surfaces, feature scaffold, and the forward-reference docs) become deterministic scripts, with a de-duplicated release gate. See release-tooling(7).
+
+## Real Execution Backend Integrations (v0.1.34)
+
+container/remote/ci backends really execute (docker/podman run, remote/CI POST-and-poll) under the sandbox contract, with byte-stable evidence vs node and fail-closed refusal when a runtime/endpoint is unavailable. See real-execution-backends(7).
+
+## Node Snapshot / Diff / Replay (v0.1.35)
+
+per-node snapshot, structural diff, and isolated deterministic replay over StateNode, reusing the v0.1.23 eval harness; fail-closed on source drift (valid|stale|absent). See node-snapshot-diff-replay(7).
+
+## Contract Migration Tooling (v0.1.36)
+
+first-class declared migration registry (run-state + workflow-app) with per-edge compatibility proofs, fail-closed reachability, and a round-trip/non-destruction prover. See contract-migration-tooling(7).
+
+## Control-Plane Scheduling (v0.1.37)
+
+priority + concurrency limits + lease lifecycle + retry/backoff + fail-closed park over the v0.1.28 Run Registry queue; policy-as-data, deterministic. See control-plane-scheduling(7).
+
+## Agent Delegation Drive (v0.1.38)
+
+spawn an external agent process per worker, capture result.md + attestation, auto-drive plan->dispatch->fulfill->accept->commit
+
+## Run Retention & Provable Reclamation (v0.1.39)
+
+tiered, append-only, cryptographically-verifiable run reclamation: seal the audit skeleton, free the reconstructable bulk, prove it
+
+## Durable State & Locking (v0.1.40)
+
+atomic temp->rename writes + fsync-durability for authoritative stores; portable stale-stealing file lock serializing the cross-process read-modify-write stores
+
+## Self-Audit Hardening & Pure-Router Decomposition (v0.1.41)
+
+evidence grounding + durable audit append + symlink-hardened containment + deterministic worker ids + recursive redaction; BackendRegistry self-describing drivers (no per-id switches); orchestrator god-object decomposed into per-domain operation modules (pure loadRun->delegate router)
+
+## Robust Result Ingest (v0.1.42)
+
+capture findings/evidence from any reasonable agent shape (alt keys + prose), CW derives grounded evidence itself, warn on empty capture — closes the v0.1.41 live-drive 'accepted with 0 captured' failure
+
+## No-False-Green Gate & Launch Prep (v0.1.43)
+
+Hard gate blocking empty-capture verifier-gated commits, plus quickstart and launch-prep docs.
+
+## Release-Gate Determinism & Agents Vendor (v0.1.44)
+
+Release-readiness checks now validate the committed blob (`git show HEAD:<path>`) instead of the mutable working tree — eliminating false-red/false-green from concurrent working-tree writes (iCloud/Spotlight/editor). Adds the `agents` vendor manifest target: a generated `.agents/plugins/cool-workflow/` adapter giving any non-Claude AI agent one common interface to CW.
+
+## P1-P2 Fixes & CI Content Surfaces (v0.1.49)
+
+Migration DAG with reversible edges (v0.1.45), capability auto-discovery (v0.1.46), vendor-adapter registry (v0.1.47), state auto-compaction and P2 fixes (v0.1.48), plus CI content-surface determinism hardening (v0.1.49).
+0.1.51
+
+0.1.76
+
+0.1.77
+
+0.1.78