musememory 3.0.2 → 3.3.0

package/.agents/artifacts/gauntlet-wave23-pr456/GAUNTLET_JOB_CONTRACT.md

@@ -0,0 +1,41 @@
+# GAUNTLET JOB CONTRACT — Wave 23 PR 4–6 (`T-460`–`T-468`)
+
+> Frozen 2026-11-08. Classification: **bounded** (three merged PR groups: agents hub, About/onboarding/deck, performance).
+> Bar: **Grafana Play** dashboard shell (live) as the external reference; the repository's own website as the internal design reference.
+> Budget: 2 rounds (hard ceiling 5). Round 1 = audit; code changes start in the fix round.
+
+## Goal
+
+Verify the Wave 23 PR 4–6 surfaces are correct, secure, responsive and honestly tested:
+
+- **T-460–T-462**: agents hub with WIRED/UNWIRED/ABSENT discovery, one-click wire/unwire (harness opt-in + config backup), point-any-agent manifest.
+- **T-463–T-465**: poster About panel with live facts, quick-start cards, searchable command deck.
+- **T-466–T-468**: projection cache with write invalidation, hidden-tab pause, sub-10ms cached projections.
+
+## Acceptance criteria (binary)
+
+1. `bun test` 0 failures; `tsc --noEmit` exit 0; `bun run build` exit 0.
+2. Headers: CSP (no wildcard), XFO DENY, nosniff, Referrer-Policy; HSTS N/A on localhost HTTP.
+3. 375 / 768 / 1280: `scrollWidth <= innerWidth`; no clipped primary controls.
+4. Agents: statuses are exactly WIRED/UNWIRED/ABSENT; the UI shows an explicit confirm before wiring and sends a harness opt-in; every config file the writer touches is backed up (timestamped) before writing; backup failure aborts; unwire removes the memory server; unknown agents fail loud on both routes.
+5. Manifest: returns the memory directory, DB path, `memory mcp`, and a snippet containing `memory_search`; no secret-shaped content.
+6. About/onboarding/deck: live facts render; four quick-start commands; seven deck commands filtered by input; every copy action is wired.
+7. Cache: second read is `X-Cache: HIT`; any mutating UI POST (or external WAL/dir change) makes the next read `MISS`; read-only POSTs stay HIT; cached bodies match.
+8. Visibility: hidden tab stops the loop; visible tab wakes it; no `setInterval`.
+9. Benchmark: median cached `/api/memories` and `/api/stats` < 10ms.
+10. No lowered bar (no skips/only/weakened assertions); `scan-secrets src` clean.
+
+## Proof commands
+
+```bash
+bun test
+bunx tsc --noEmit
+bun run build
+bun bin/musememory.ts scan-secrets src
+curl -sI http://127.0.0.1:3899/ | rg -i '(content-security-policy|x-frame-options|x-content-type-options|referrer-policy)'
+# viewport + shell audit via the Playwright script used in PR 1
+```
+
+## Scoring
+
+`S = 0.40·Correctness + 0.25·MinimalDiff + 0.20·EdgeCases + 0.15·Cleanliness`. Pass: `S >= 9.0`, zero blockers, all criteria true.

package/.agents/artifacts/gauntlet-wave23-pr456/ITERATION_LEDGER.md

@@ -0,0 +1,36 @@
+# ITERATION LEDGER — Wave 23 PR 4–6
+
+> Contract: `GAUNTLET_JOB_CONTRACT.md`. Budget 2 rounds; Round 1 audit, fix round applied.
+
+## Round 1 — audit (2026-11-08)
+
+- **Gate**: green. 1471/0, tsc clean, build 434 modules, headers 4/4, 375/768/1280 no overflow, cache MISS→HIT, secret scan clean.
+- **Fresh Critic: 7.35/10 — FAIL**, criterion 4 false:
+  - **B1** wire backed up only `detected.configPath`; `claude-json` also rewrites `~/.claude/settings.json` with no backup; backup failures were swallowed; fixed `.bak` name overwrote the pristine copy; `id:"all"` wrote every agent with zero backups; XDG divergence.
+  - **B2** projection cache invalidated only by in-process POSTs, so external writes were invisible until a UI POST; read-only POSTs (`/api/search`, `/api/cognition/why`) invalidated unnecessarily.
+  - **B3** unwire unknown agent returned 200 `success:false`; the client flashed success on any 200; harness opt-in was a silent client constant.
+  - **B4** 44×44 claim was contradicted by bare `.btn-icon` copy buttons and a 24px chip override.
+- Non-blocking: agents detection on every load (sync `$HOME` sweep), source-grep-only tests, missing count assertions.
+
+## Fix round (branch `fix/wave23-gauntlet`)
+
+- Wire: rejects `id:"all"`; typed unknown-agent 404 for wire and unwire; backs up every target the write can touch (`~/.claude.json` + `~/.claude/settings.json`, XDG-aware via the detector path); timestamped backups that never overwrite; backup failure aborts the request with 500; response carries the backup paths.
+- Cache: key includes directory + DB mtime freshness, so external writes miss; version bump restricted to an explicit mutating-route allowlist.
+- Client: explicit confirm before wiring; only flashes success when `data.success === true`; failure shows the report message; backup path surfaced.
+- CSS: 44px applies to every phone control (`btn-icon`, chips, buttons, inputs), 24px chip override removed.
+
+**Verification**: `bun test` **1471 pass / 0 fail**, tsc clean, build 434 modules, secret scan clean, cache MISS→HIT, 375/768 targets 44×44.
+
+Status: **CAPPED** — fixed after the only critic round (budget 2); no second critic. Code review follows.
+
+## Code review — diff mode (Linus style)
+
+Verdict: **REQUEST-CHANGES** on the gauntlet fix commit. Findings and resolution:
+
+- **CR1 (blocking)** `freshnessToken` probed only `<memoryDir>` and a hardcoded `memory.db`, so WAL commits (`memory.db-wal`) and the `memories/` mirror were invisible; a failed stat collapsed to `0`. Fixed: probe `dir`, `getStorageLayout(dir).memoriesDir`, `dbPath`, `dbPath-wal`, `dbPath-journal`; a missing path contributes `"-"`, never a fake zero.
+- **CR2 (blocking)** neither new cache behavior had a test, and the fail-loud backup path was unverified. Added: read-only POST (`/api/cognition/why`) stays `HIT`; an external `propose` through a second open store yields `MISS` and the new memory; an unwritable backup directory returns 500 and leaves the config without a memory server.
+- **Doc drift** the contract said "any UI POST" while the packet scored a mutation allowlist. Both now say: mutating POST or external WAL/dir change → `MISS`; read-only POSTs stay `HIT`. The harness row now names the confirm dialog as the consent surface.
+- **Accepted nitpicks**: invalidation inverted to a read-only set (forgotten routes miss, never stale), backup names carry a monotonic counter, all backup paths are surfaced to the user.
+- Noted, not changed in this branch: unwire leaves `allowedTools` grants; `.bak` files accumulate; mesh projections have no freshness signal.
+
+**Post-review gate**: `bun test` 1471/0, tsc clean, build 434 modules, secret scan clean.

package/.agents/artifacts/gauntlet-wave24-25/ACCEPTANCE_PACKET.md

@@ -0,0 +1,13 @@
+# ACCEPTANCE PACKET — Waves 24–25
+
+**Status: `CAPPED`** — the diff review returned REJECT and all blocking findings were fixed and re-gated; there was no second review round. Website visual verification is **N/A** because the repository's Astro prerender fails on its own `cookie` dependency before page generation.
+
+## Receipts
+
+- `bun test` 1488 / 0 (226 files)
+- `tsc --noEmit` clean
+- `bun run build` 434 modules
+- `scan-secrets src` clean
+- Playbook safety rails survive MCP step-only updates; corrupt store fails loud and stays byte-identical; episode ids unique; bi-temporal offsets parsed; CLI honours `--dir`.
+
+Residual debt: `TECHNICAL_DEBT-wave24-25.md` (wiring episodes/temporal/decay, Astro cookie interop, playbook execution semantics).

package/.agents/artifacts/gauntlet-wave24-25/ITERATION_LEDGER.md

@@ -0,0 +1,28 @@
+# ITERATION LEDGER — Waves 24–25 (2026-11-08)
+
+## Automated gate
+
+- `bun test`: 1488 pass / 0 fail (226 files)
+- `tsc --noEmit`: clean
+- `bun run build`: 434 modules
+- `scan-secrets src`: clean
+- Website visual gate: **N/A** — `astro build` fails in Astro's own prerender import (`cookie` CommonJS named export `parseCookie`) before page generation; Vite entrypoints compile. Not caused by this diff; blocks rendering `agents.astro` and `playground.astro`.
+
+## Diff review (Linus style) — REJECT, then fixed
+
+Findings and resolution:
+- **B1** MCP updates synthesized `[]` for absent arrays, erasing playbook safety rails. Fixed: arrays are spread only when present; MCP resolves title→id before updates; test asserts rails survive a step-only update.
+- **B2** corrupt `playbooks.json` read as empty and was overwritten. Fixed: `PlaybookStoreError` on invalid/non-array JSON; CLI returns 1; MCP returns a tool error; test asserts the file is left byte-identical.
+- **RC1** CLI printed the pre-outcome snapshot. Fixed: print the updated record.
+- **RC2** title was the write identity. Fixed: updates key on `id` (with title→id resolution at the MCP boundary); same-titled playbooks coexist.
+- **RC3** CLI ignored `--dir`/`--global` and treated `--run` as list. Fixed: dir/global resolution, `--run`/`--show` aliases.
+- **RC4/RC5** episode ids collided within a second; `test` preempted goal. Fixed: per-session sequence ids; kind-first and anchored phase patterns.
+- **RC8** timestamps compared as strings. Fixed: epoch parsing, invalid timestamps throw, `ended_at` fallback removed.
+- **RC9** step shape/goal unvalidated. Fixed: non-empty goal and per-step action validation.
+- **RC10** benchmark page commands/wording were wrong. Fixed: `claims:check`, real CLI benchmark command, single-value claims semantics, descope note.
+- **RC11** playground claimed schema parity. Fixed: copy says four-column subset.
+- **RC6/RC7** `episodes`, `temporal`, `decay` have no production callers. Documented as scaffolding in `TECHNICAL_DEBT-wave24-25.md` and ticket-plan notes.
+
+## Status
+
+**CAPPED** — fixed after the single review round; no second reviewer. Visual/a11y gate remains blocked by the Astro prerender issue.

package/.agents/artifacts/gauntlet-wave25-wiring/ACCEPTANCE_PACKET.md

@@ -0,0 +1,7 @@
+# ACCEPTANCE PACKET — site build + Wave 25 wiring
+
+**Status: CAPPED.** Audit mode returned REQUEST-CHANGES; every blocking finding was fixed and re-gated. No second audit round.
+
+Receipts: tests 1487/0; tsc clean; build 434 modules; secret scan clean; site build 6 pages; site QA 17/17 SHIP.
+
+Residual: failure observations (`source: test`) still classify as verification in episodes; `asserted_at` has no SQL column/index (JS gate only); no second audit.

package/.agents/artifacts/gauntlet-wave25-wiring/ITERATION_LEDGER.md

@@ -0,0 +1,27 @@
+# ITERATION LEDGER — site build + Wave 25 wiring audit (2026-11-08)
+
+## Audit (code-review, audit mode) — REQUEST-CHANGES
+
+| Finding | Resolution |
+|---|---|
+| B1 global `cookie` override rewrote express's dependency and fixed nothing | Override removed; lock regenerated |
+| B2 `patch-astro-cookie.mjs` was cargo-cult and could mask a bad resolution | Script and call sites deleted; Node launcher alone builds |
+| R3 `asserted_at` had readers and no writer | `propose` accepts `asserted_at`; `memory_write` schema declares it; persistence tested |
+| R4 two absent-value semantics | `temporal.ts` and `retrieval.ts` now both treat absent as "no belief constraint" |
+| R5 briefing test used type-illegal observation sources | Fixtures use `manual`/`test`; failure provenance still classified as verification (debt) |
+| R6 briefing truncated observations before episodes | Episodes built over the full stream, then sliced |
+| R7 "wired" claims were one shared string | Ticket/debt notes now state per-ticket facts |
+| D6 site prerender open in debt | Struck as fixed |
+
+Also updated `.agents/standards/frontend-astro.md` (Astro 7.4.7 + Node launcher note).
+
+## Gate
+
+- `bun test`: 1487 pass / 0 fail (227 files)
+- `tsc --noEmit`: clean
+- `bun run build`: 434 modules
+- `scan-secrets src`: clean
+- `node scripts/astro.mjs build`: 6 pages
+- `site:qa`: 17/17 passed, VERDICT SHIP
+
+Status: **CAPPED** — fixes applied after the audit; no second audit round.

package/.agents/artifacts/gauntlet-wave27/ACCEPTANCE_PACKET.md

@@ -0,0 +1,7 @@
+# ACCEPTANCE PACKET — Wave 27 hardening
+
+**Status: CAPPED.** Code review and muse-security both rejected the first cut; blockers were fixed and re-gated. No second critic round.
+
+Receipts: 1521/0 tests; tsc clean; build 434 modules; secret scan clean. Crypto: GCM AAD + key id, Ed25519 domain-separated signatures, strict tenant/branch names, 0600/0700 files, overlay merge through the three-way resolver with audit.
+
+Residual: `TECHNICAL_DEBT-wave27.md` (chain truncation/TOFU, key rotation, semantic merge poisoning, scope:client not wired, overlay GC, branch identity from store.dir).

package/.agents/artifacts/gauntlet-wave27/ITERATION_LEDGER.md

@@ -0,0 +1,36 @@
+# ITERATION LEDGER — Wave 27 (2026-11-09)
+
+## Critics
+
+- **code-review audit / gauntlet**: REJECT, 4.5/10 (Correctness 4, Diff 6, Edge 4, Architecture 4).
+- **muse-security audit**: 8 High, 9 Medium, 0 Critical; headline claims overstated.
+
+## Fixed in this round
+
+| Finding | Fix |
+|---|---|
+| Enclave `safeName` collisions | Strict `[A-Za-z0-9._-]` validation; reject instead of rewrite |
+| Envelope metadata not authenticated | GCM AAD over version/key_id/client/division; request metadata asserted |
+| Master-key TOCTOU | `wx` creation + EEXIST re-read; 64-hex env only; `key_id` in envelope/AAD |
+| Enclave path traversal (`..`) | Name validation + containment by construction |
+| Enclave permissions | 0700 dirs, 0600 envelopes; key 0600 |
+| Branch name collisions | Reversible base64url filenames; `meta.branch` identity check |
+| Overlay path collision with memory versioning | Moved to `.memory/overlays/` |
+| Overlay merge lost updates | Captured `base_entries`; merge now calls `mergeMemories3Way` (T-439 wiring) |
+| Overlay merge non-atomic / unaudited | Whole-batch validation + `withStoreLock` + per-entry `overlay_merge` audit + `archiveMemory` for deletions |
+| Overlay CLI had no write path | `overlay set` / `overlay rm` + audit |
+| Detached HEAD treated as branch "HEAD" | `git symbolic-ref`; null on detached, requires `--branch` |
+| `overlay list` ignored `--branch` | Uses resolved branch |
+| Attest keygen unsafe default/overwrite | Defaults to `~/.config/musememory/keys` (0700), refuses overwrite unless `--force`, audit event |
+| Canonical JSON collisions | Rejects unsafe numbers, non-plain objects; omits undefined keys; domain separator |
+| Chain semantics | Empty chain rejected; version/algorithm signed and checked; `verify-chain` CLI |
+| merge3 fail-open/manual contract | Exit 1 on conflicts unless `--allow-conflicts`; TTY guard; stderr prompts; 0600 output; secret scan |
+| Enclave CLI type/traversal/content | Enum validation, name validation, `--file`/stdin |
+
+## Gate
+
+- `bun test` 1521 pass / 0 fail (240 files)
+- `tsc --noEmit` clean; build 434 modules; secret scan clean
+- New regression tests: name collisions, AAD relabel, empty/tampered chains, overlay 3-way conflict, CLI fail-closed, overlay write surface.
+
+Status: **CAPPED** — one critic round; residual items in `TECHNICAL_DEBT-wave27.md`.

package/.agents/context/current.md

@@ -1,7 +1,46 @@
 # 📍 Current Shipped State & Active Invariants
 
-> **Version**: `v3.0.2`  
-> **Status**: Patch Release — correctness fixes from the cross-file invariant audit, machine-verified published claims, and the website redesign.
+> **Version**: `v3.1.3`  
+> **Status**: Wave 29 Landed — Sovereign LifeOS Migration & Complete Decommissioning (T-444 to T-448).
+
+---
+
+## 🧾 Session — Ticket Ledger Renumber & Remainder Audit (2026-11-08)
+
+Wave 23's `T-405`–`T-422` collided with the already-landed `T-405` (conflicted injection) and `T-406` (project casing), so that block is now `T-451`–`T-468` across the plan, roadmap and this file. The historical blueprint keeps its original IDs with a translation note; `HANDOFF.md`'s `T-405`/`T-406` are the unrelated landed fixes and were left untouched. Every unchecked ticket was then verified against source: 4 were already implemented (`T-150`, `T-151`, `T-153`, `T-384`) and are now marked done, 15 are partial, 86 are genuinely unbuilt, `T-163` is dropped by principal decision D3, and `T-449`/`T-450` remain filed/deferred. Full evidence table: `.agents/context/ticket-plan.md` → v15. Doc-claim gate re-run green (20/20).
+
+### Wave 23 PR 1 implemented (local, not landed)
+
+`T-451`–`T-453` are implemented in the new `src/ui_assets.ts` module and wired through `src/ui.ts`:
+
+- **T-451** — the studio now renders a fixed 76px left rail with six numbered sections (`01 OVERVIEW` … `06 ABOUT`), a tools group for the pre-existing views, a warm `#0c0a09` palette, and Archivo Black / JetBrains Mono typography. The rail uses words and numbers, not emoji.
+- **T-452** — the header carries a project / global / mesh segmented switcher that refetches live. `/api/memories`, `/api/stats`, `/api/current`, `/api/wiki`, `/api/user-profile` and `/api/audit` resolve the scope; the global store (`~/.memory`) opens lazily and is never created by a read; mesh results carry `origin_project`. Mutations stay project-scoped and the studio shows a scope banner whenever it is not viewing the project store.
+- **T-453** — CSS, markup and client JS are split out of the old 2,700-line template literal and served from `/assets/studio.css` and `/assets/studio.js` with strong ETags, pre-gzipped buffers, `Vary: Accept-Encoding` and 304 fast paths. The standalone export inlines the same modules.
+
+Gate at merge: `bun test` **1451 pass / 0 fail** across 217 files, `tsc --noEmit` clean, `bun run build` 434 modules. **Landed in [PR #337](https://github.com/harshsinghmp/musememory/pull/337) at `d0c9600`** (2026-11-08); `check:landed` → `RESULT: LANDED`.
+
+### Gauntlet run — CAPPED at the 5-round ceiling (2026-11-08)
+
+Ran the bounded gauntlet loop against this PR (contract + ledger + packet in `.agents/artifacts/gauntlet-wave23-pr1/`). Five rounds: Fresh Critic scores **5.95 → 7.63 → 7.90 → 8.85 → 7.99**. Both R5 blockers were confirmed closed (mesh-link cache invalidation; scope-generation guards on every scoped fetch); the final critic left only small markup/ARIA/banner defects, which were fixed in the post-ceiling Integrator delta and verified by the gate but not re-critiqued. The packet is explicitly **CAPPED, not a PASS**. Final gate: **1450/0**, tsc clean, build 434 modules, zero overflow at 375/768/1280, 44×44 targets at 375/768, headers 4/4 (HSTS N/A), per-encoding ETags + wildcard 304 + 406 negotiation, secret scan clean. Residual debt (scope-race test, CSP `unsafe-inline`, mesh read-only store minting, duplicated `escapeHtml`, standalone export zeros) is in `TECHNICAL_DEBT.md`.
+
+---
+
+## 🔧 Session — Sync Recovery (2026-11-03)
+
+Restored `AGENTS.md` byte-exact from HEAD after an `updateagents` run from the umbrella directory replaced it (and persona files in sibling repos) with three-line adapters. Added `.commandcode/` and `.stversions/` to `.gitignore`, made the Syncthing ignore rules canonical in `~/.dotfiles/syncthing/`, and recorded the workspace-isolation invariant. The engine fix now stops at nested repository roots and refuses umbrella directories; the global install was refreshed. Repo-local recovery for `muse-agents` and `muse-skills` is owned by their respective agents.
+
+---
+
+## 🏛️ Wave 29 — Sovereign Ecosystem Migration & Decommissioning Complete (2026-10-26)
+
+All personal identity, telos, operational rules, and hardware terrain scanning migrated into sovereign `musememory` engines, native Vibeguard secret scanning, and de-LifeOSified agent governance. Legacy import tools have been permanently purged from source:
+
+- **Sovereign Persona & Store Ingestion (`T-444`)**: Identity, telos, and operational rules synthesized directly into sovereign `~/.memory/USER.md` persona and SQLite store. Legacy tool `import_lifeos.ts` permanently deleted from source.
+- **Native Vibeguard CLI (`T-445`)**: `memory scan-secrets [dir]` standalone zero-dependency secret scanner replacing legacy external scripts (`test/cli_secrets.test.ts`).
+- **Machine Terrain Collector (`T-446`)**: `memory terrain [--out file.md]` compiles deterministic hardware, toolchain, and local agent ecosystem maps into `system.md` (`test/vault_system.test.ts`).
+- **Global Governance De-LifeOSification (`T-447`)**: Removed all 6 legacy imports and paths from `~/.agents/AGENTS.md` and repository standards, binding truth directly to `~/.memory/USER.md` and `memory` MCP.
+- **Sovereignty Decommissioning Gate (`T-448`)**: Enforces zero external legacy imports, confirms legacy tool removal, and passes 100% cleanly in `test/sovereignty_gate.test.ts`.
+- **Golden Sequence Documentation**: Published the complete 4-phase best-practice sequence for using Muse Memory with AI coding agents in `README.md`, `docs/agent-reference.md`, and `docs/cli-reference.md`.
 
 ---
 
@@ -125,15 +164,113 @@ on any blocking failure, so CI runs it. Report:
 
 ---
 
-## ⏭️ Next Up (Post-Wave 22 Focus)
-- **Production-Grade Showcase Demo**: Re-record high-fidelity terminal demo video/GIF reflecting v3.0.0 capabilities (failure attribution, HOT.md compounding, memory branch/checkout, MCP context recall) to replace removed legacy demo.
-- **Wave 23**: Review next backlog items from `.agents/context/ticket-plan.md` (e.g. Hebbian co-activation retrieval expansion, query-seeded PPR restart set, retrieval weight auto-tuner, and contradiction engine calibration).
-- **Consolidate the remaining duplicated tables** (deferred from this session as behaviour-changing; full detail with exact locations in [`docs/roadmap.md`](../../docs/roadmap.md) → *Engineering debt*): the three divergent per-type TTL tables in `src/retention.ts`, the eight separate token estimators, the divergent limit-clamping ceilings across retrieval paths, and the remaining zero-caller exports.
+## ⏭️ Next Up (Post-Release Focus)
+- **Waves 23–25 are complete.** Next is **Wave 26–29** (`T-432`–`T-443`): AST associative retrieval (`T-432`–`T-435`), branchable/enclave memory (`T-436`–`T-439`), test-driven reinforcement (`T-440`–`T-443`). Site note: unblock the Astro/cookie prerender error before any visual QA on the new pages:
+  - Full UI overhaul inspired by Ryoku Rashin (`http://127.0.0.1:3600`): fixed 76px left-rail cyber-minimalist navigation, instant Canvas Graph ↔ Filterable List toggle, 60fps kinetic-settling physics loop, and 52-week activity calendar heatmap.
+  - Multi-scope switcher: Project Scope (`./.memory`) vs Global User Scope (`~/.memory`) vs Cross-Project Mesh.
+  - Workspace Terrain Maps (`stack.md`, `habits.md`, `ownership.md`, `packages.md`, `repo.md`, `logs.md`) for token-efficient agent orientation.
+  - Universal Agents Hub: Auto-discovery, 1-click MCP wiring, and drop-in prompt snippets for 10+ agents (Antigravity, OpenCode, Claude Code, Cursor, Copilot, Hermes).
+  - Poster-style About panel with live architectural facts, interactive quick start onboarding, and CLI command deck.
+  - Sub-10ms UI serving performance with in-memory SQLite read caching and visibility-based zero-CPU idling.
+- **Production-Grade Showcase Demo**: Re-record high-fidelity terminal demo video/GIF reflecting v3.0.0 capabilities to replace removed legacy demo.
+- **Wave 24 remainder**: `T-316` (Hebbian co-activation edges into write-time `linked_memory_ids` seeding — **unblocked**, dep `T-095` landed; note the ticket premise is partly stale: `T-307` already landed the retrieval-expansion side) and `T-318` (retrieval weight auto-tuner from ROI telemetry — **unblocked**, deps `T-127` + `T-140` both landed). `T-317` (query-seeded PPR restart set) stays **blocked** on `T-178` (graph retrieval A/B gate).
+- **Consolidate the remaining duplicated tables** (deferred from this session as behaviour-changing; full detail with exact locations in [`docs/roadmap.md`](../../docs/roadmap.md) → *Engineering debt*): the three divergent per-type TTL tables (`src/retention.ts` **and** `src/retention/expiry_policy.ts` — only two of the three are in `retention.ts`), the **nine** separate token estimators (across nine files, plus six inline expressions — the collapse is *not* behaviour-neutral, because `src/orchestrator/context.ts` deliberately consumes the pessimistic measure), the divergent limit-clamping ceilings across retrieval paths (**seven** sites in `src/mcp.ts` alone), and the remaining zero-caller exports (twelve genuinely dead; 74 exported with no external caller).
 - **Remove the legacy `web/` directory** — done in this session. It was a second, hand-written copy of the site that `apps/website/` superseded; keeping it was the same "second copy that gets forgotten" failure mode this session fixed everywhere else.
 
 ---
 
 ## 🏁 Recently Landed Waves (Verified on `dev`)
+- [x] **Outcome pilot + RC blockers** (2026-11-09) — with/without pilot 2/2 vs 1/2; latency claims corrected to p95 targets; T-450 ADR-008 boundary; overlay GC, branch root identity, enclave put lock.
+- [x] **Evidence + key lifecycle** (2026-11-09) — dogfood benchmark (hit@3 10/10, MRR 0.933, 68% token savings), evidence ledger with 2 quarantined claims, benchmark title-less crash fixed, `enclave keys/prune`, backup drill, runbook.
+- [x] **Enclave key rotation + client-scope descope** (2026-11-09) — versioned keyring, `enclave rotate`, old-key rollback; `scope: client` descoped per ADR-007.
+- [x] **Attestation v2 + merge guard** (2026-11-09) — seq/head/key-id pinning, truncation rejection; semantic merge blocks untrusted lineage, inflation and protected constraints.
+- [x] **Wave 27 hardening** (2026-11-09) — audit-REJECT fixes: strict names, GCM AAD/key id, audited overlay merge via T-439, fail-closed merge3, corrected claims. Gate 1521/0; CAPPED.
+- [x] **Wave 27 (`T-436`–`T-439`)** (2026-11-09) — branch overlays (#377), encrypted enclaves (#378), packet attestations (#379), three-way merge (#380). Tests 1514/0.
+- [x] **Sensitive-read audit + project identity** (2026-11-09) — `read` audit events for `scope: "all"`/untrusted; configured project via env/config/`memory project`.
+- [x] **Code-context security boundary** (2026-11-09) — project-scoped by default, untrusted opt-in, read-time redaction/sanitization, bounded inputs; `muse_memory_for_code` accepts project/include_untrusted.
+- [x] **Unified code resolver** (2026-11-09) — `resolveMemoriesForCode` is the single matcher; pre-flight and PR context adapt it.
+- [x] **Pre-flight surfaces** (2026-11-09) — PreToolUse hook + read-only `muse_preflight` MCP tool; 119 tools.
+- [x] **Wave 26 (`T-432`–`T-435`)** (2026-11-09) — recursive-CTE associative expansion (`T-432`, direct `dev`), pre-flight context (`T-433`, #370), blast-radius risk (`T-434`, #371), anchor-matched auto-recall (`T-435`, #372). Tests 1495/0.
+- [x] **Site build + scaffolding wiring** (2026-11-08) — PR #367 (Node Astro launcher; site:qa 17/17 SHIP) and #368 (briefing episodes, explicit `asserted_at` belief clock in `as_of`, duplicate decay removed).
+- [x] **Wave 24–25 review hardening** (2026-11-08) — PR #366 merged at `a485875`; playbook rail preservation, fail-loud corrupt store, id-keyed updates, unique episode ids, goal-first phases, epoch timestamps, CLI flags, doc honesty. Gauntlet: REJECT → fixed, CAPPED; website visual gate blocked by pre-existing Astro/cookie prerender error.
+- [x] **Wave 24 (`T-423`–`T-426`)** (2026-11-08) — terminal demo script/recipe, vendored sql.js playground, capability-only benchmark matrix, accessible agent demos. PRs #362–#365. Site prerender has a pre-existing Astro/cookie CJS blocker (entrypoints compile).
+- [x] **Wave 25 (`T-427`–`T-431`)** (2026-11-08) — playbooks, episodes, bi-temporal belief windows, reinforced decay, and playbook CLI/MCP commands. PRs #356–#361.
+- [x] **Post-Wave-23 debt sweep** (2026-11-08) — PR #355 merged at `432cf01`: agents discovery cache, backup retention, unwire tool-grant cleanup, mesh freshness bucket.
+- [x] **Wave 23 gauntlet hardening** (2026-11-08) — PR #354 merged at `d2ea59c`; complete config backups (XDG-aware, timestamped, fail-loud), unknown-agent 404s, honest wire reporting, WAL/layout cache freshness, 44px phone controls. Gauntlet: 7.35 critic → REQUEST-CHANGES review → fixed; **CAPPED, not a PASS**. Evidence: `.agents/artifacts/gauntlet-wave23-pr456/`.
+- [x] **Wave 23 PR 6 (`T-466`–`T-468`)** (2026-11-08) — projection cache, hidden-tab pause and the UI performance benchmark. **Wave 23 complete.** PRs #351–#353.
+- [x] **Wave 23 PR 5 (`T-463`–`T-465`)** (2026-11-08) — poster About panel, quick-start cards and searchable command deck. PRs #348–#350.
+- [x] **Wave 23 PR 4 (`T-460`–`T-462`)** (2026-11-08) — agents hub: discovery statuses, one-click wire/unwire with `.bak` backups and harness opt-in, point-any-agent manifest. PRs #345–#347.
+- [x] **Wave 23 PR 3 / `T-459`** (2026-11-08) — PR #344 merged; generation-time redaction plus a token-savings and leak test (`test/vault_terrain.test.ts`).
+- [x] **Wave 23 PR 3 / `T-458`** (2026-11-08) — PR #343 merged at `65113af`; terrain vault split-pane (tree, GEN badges, markdown viewer, REINDEX scanline).
+- [x] **Wave 23 PR 3 / `T-457`** (2026-11-08) — PR #342 merged at `8006427`; deterministic workspace terrain generator and `terrain --workspace` CLI.
+- [x] **Wave 23 PR 2 / `T-455`** (2026-11-08) — PR #340 merged at `05b4fdf`; energy-settling loop with wake hooks, deterministic ring seed, pointer capture.
+- [x] **Wave 23 PR 2 / `T-456`** (2026-11-08) — PR #341 merged at `436d198`; scope-aware activity endpoint, 52-week heatmap and agent session cards.
+- [x] **T-449 test hermeticity** (2026-11-08) — PR #338 merged at `94620895`; `findProjectRoot` no longer adopts `$HOME`/`/` as a project root, doc-link scan skips local sync state, regression test with a fake home. Earlier caller-scan fix merged via #337 (`c0f7aef`). Issue #336 closes.
+- [x] **Wave 23 PR 2 / `T-454`** (2026-11-08) — PR #339 merged at `3e621fa`; instant graph/list toggle with a sortable table, `aria-sort`, and a 100ms debounced filter.
+- [x] **Wave 23 PR 1 (`T-451`–`T-453`)** (2026-11-08) — merged in PR #337 at `d0c9600`; `check:landed` → **LANDED**. Left rail, dual-scope switcher, modular ETag/gzip assets, read-only global/mesh reads, 375/768/1280 verified, 1451/0 at merge. Follow-ups in PR #337's thread: stateful-regex flake fixed (`c0f7aef`), npm Bun CI fallback (`b298beb`). Gauntlet verdict: CAPPED at 7.99 (not a 9.0 PASS) — evidence in `.agents/artifacts/gauntlet-wave23-pr1/`.
+- [x] **Released: v3.1.2** (2026-10-25) — completes the silent-casing fan-out:
+  - PR #330 merged at `d7e875b` (*"Release: v3.1.2"*); annotated tag
+    [`v3.1.2`](https://github.com/harshsinghmp/musememory/releases/tag/v3.1.2)
+    pushed on `d7e875b`; GitHub release published. Back-merge #331 merged at
+    `fb8ff75`; `dev` re-synced.
+  - `package.json` is `3.1.2`; `[Unreleased]` empty.
+  - npm publish still blocked (same E404 as v3.1.0/v3.1.1 — principal-only).
+
+- [x] **Released: v3.1.1** (2026-10-24) — the two post-release defect fixes shipped to `main`:
+  - PR #324 merged at `5f97870` (*"Release: v3.1.1"*); annotated tag
+    [`v3.1.1`](https://github.com/harshsinghmp/musememory/releases/tag/v3.1.1)
+    pushed on the release merge; GitHub release published with notes derived
+    from the stamped changelog. `package.json` is `3.1.1` and `[Unreleased]`
+    is empty again.
+  - Governance disclosure: the `main` branch-guard ruleset requires one
+    approving review; the merge used `--admin` on the principal's order to
+    cut the release (the `test` check was green on the release branch first:
+    1408 / 0 across 213 files).
+  - Back-merge reunited `main` with `dev` (PR #325; `dev` push CI green,
+    3·m·12s). Post-merge `check:landed` → **RESULT: LANDED** at `ebf4040`.
+  - Standing blocker disclosed again: the `npm-publish` workflow failed on the
+    tag with the same E404 as v3.1.0's tag (query history: v3.0.1/v3.0.2
+    published fine on Oct 22; v3.1.0 + v3.1.1 both 404 on `PUT musememory@…`,
+    so the CI npm token lost owner rights to the package between Oct 22 and
+    Oct 24). npm publishing stays **principal-only** (owner credentials + 2FA)
+    until the token is repaired.
+  - CI false positive filed: the #325 back-merge PR's delivery-gate step
+    failed `check-landed --pr --branch main` because the gate greps for an
+    *open* PR at the head branch and the run's queue lag let the merge close
+    it first (`RESULT: NOT REVIEWABLE` on identical, separately-green
+    content). Fence in the gate's head-branch semantics for back-merge PRs.
+- [x] **Fix trains: T-405 + T-406** (2026-10-24) — issues #313/#314, delivered through the gauntlet-loop + code-review engines (contract-frozen scope, blind Fresh Critic, mutation-verified tests):
+  - **T-405** ([#313](https://github.com/harshsinghmp/musememory/issues/313)): PR #321 merged at `e30a1b7`. Conflicted memories are now excluded **by default** from every prompt-serving retrieval path — queryContext lifecycle gate (hidden_count-aware, status-scoped-guarded), FTS store gate with the overfetch connective armed iff any post-fetch hide-gate stays active, vector row gate + engine passthrough, ranking candidate veto before scoring, muse-context pool choke point + anchor/route lens strips + constraint projection + withheld notice — all sharing one `includeConflicted` opt-in (`SearchOptions` cascade; MCP `include_conflicted` schema symmetry on both search/memory_recall tools; CLI `--include-conflicted`). The frozen golden recall fixture was deliberately untouched — the eval corpus carries zero conflicted entries (verified, not assumed); new tests use a self-contained corpus (T-143 promotion discipline respected). RED receipts real (conflicted entry served `kept:high_relevance` into `confident_answers`); critic verdict **9.5/10** with a mutation-verified overfetch test. As-of/temporal queries deliberately keep serving state-at-the-time (documented decision).
+  - **T-406** ([#314](https://github.com/harshsinghmp/musememory/issues/314)): PR #322 merged at `d3a8f6b`. Project names resolve case-insensitively (`resolveProjectName`/`listProjectNames`/`formatAvailable`, `src/health/resolve.ts`) on all three health surfaces, failing **loud** (CLI exit 1 · MCP isError · Studio 404) with the requested name echoed and the available list; resolved-with-case-diff surfaces the canonical label plus a note naming the requested form; `--json` stdout stays pure JSON (probe-verified). The gate stays a pure exact-canonical reporter — `git diff test/health_project_scope.test.ts` empty. Final critic verdict **9/10**.
+  - Full-suite receipts at merge time: `bun test` **1408 pass / 0 fail** (213 files, +13 tests); `tsc --noEmit` clean; claims `1400→1408`; docs links all resolve; secret scan clean; `check:landed:pr` → `RESULT: REVIEWABLE` before both merges, `check:landed` → **`RESULT: LANDED`** after (at `e30a1b7`, then at `d3a8f6b`).
+  - **T-406 fan-out** ([#327](https://github.com/harshsinghmp/musememory/issues/327)): PR #328 merged at `17c0abe`. Same silent-casing defect closed on the remaining 7 audit surfaces — MCP `memory_anchor_audit`, `memory_drift_audit`, `muse_bug_clusters`, `muse_tech_debt` + studio `/api/cognition/anchors`, `/api/cognition/debt`, `/api/cognition/clusters` — all routed through the existing `resolveProjectName()` owner; unknown projects now fail loud (MCP tool error · Studio 404) with the available list. Gates/engines stay exact. +17 tests (1425 total, 0 new failures vs the 9 pre-existing env failures on dev).
+  - Filed follow-ups (deliberate ledger, see HANDOFF): ambiguous-case wording, Studio client 404 echo, silent-casing fan-out (`muse_tech_debt`, anchor/drift audits, `/api/drift`, `/api/cognition/*`), wikilink hop-gate conflicted filter, `get_context` include_superseded schema-vs-handler mismatch, hybrid-path status plumbing, queryContext superseded-gate status guard.
+- [x] **Released: v3.1.0** (2026-10-24) — Waves 23–26 shipped to `main`:
+  - PR #318 merged at `c5a141d` (*"Release: v3.1.0"*); annotated tag
+    [`v3.1.0`](https://github.com/harshsinghmp/musememory/releases/tag/v3.1.0)
+    pushed on the release commit; GitHub release published with notes
+    derived from the stamped changelog. `package.json` is `3.1.0` and
+    `[Unreleased]` is empty again.
+  - Governance disclosure: the `main` branch-guard ruleset requires one
+    approving review; the merge used `--admin` on the principal's order to
+    cut the release (the `test` check was green first). Disclosed also in
+    the handoff.
+  - The post-release back-merge reunited `main` with `dev`
+    (`fix/showreel-hash-path` lived only on `main` until now; the branch is
+    pruned). Gate evidence, quoted: `bun run check:landed` →
+    **RESULT: LANDED**; `git merge-base --is-ancestor HEAD origin/dev`
+    exits 0 at `c5a141d`; `bun test` re-run on the merged tree: 1393 / 0.
+- [x] **Wave 26: The Landing Gate — a commit is not a delivery**:
+  - On 2026-10-22 two complete waves (T-320, T-321) were committed to a local branch and a local `dev`, and the session ended. No push, no PR, no CI run — while the close-out document described them as landed, under a heading reading *"Recently Landed Waves (Verified on `dev`)"*. `dev` did not contain them; `origin/dev` had already moved to v3.0.2, so the branch had diverged and the work was invisible to review and to the release. **The code was correct and every test was green. The failure was a false claim about the remote that nothing had ever checked.**
+  - `scripts/check-landed.ts` is now the machine-checkable form of that claim: work is **landed** only when it is reachable from the integration branch **on the remote**, and until then it is **local**. The gate verifies `HEAD` is an ancestor of `origin/dev` and that nothing is unpushed, exits non-zero otherwise, and prints the exact commands and SHAs so a report can quote the evidence rather than assert it. A pre-review mode (`--pr --branch <name>`) verifies a branch is pushed, has an open PR, and has no failing checks. Wired into CI for both `pull_request` and `push`.
+  - `.agents/standards/landing-gate.md` (routed from `AGENTS.md` as core invariant #7) restricts the vocabulary — *landed / shipped / merged / released / "verified on `dev`"* may only be used after the gate exits 0 — and requires every final report to carry a Remote State block. The closeout DOX pass runs the gate before writing under a "Landed Waves" heading.
+  - `.agents/artifacts/WORKTREE-LEASE.md` — the file was referenced by an earlier handoff but did not exist. It now defines the shared-checkout lease (taken before git mutations, released **after the push**, not the commit) and the staging rule: **never `git add -A` in a checkout another agent may be writing.** A broad `git add` silently absorbs a concurrent agent's in-progress edits into your commit, producing a commit whose message describes work that is not in it — which happened for real on 2026-10-22 and was caught only because a subagent reported it.
+- [x] **Wave 25: T-321 Surface Completion**:
+  - A verification audit of the Wave 24 close-out found the *functions* took `options.project` but five *surfaces* never plumbed it: `memory_anchor_audit` and `memory_drift_audit` (MCP), and `/api/drift`, `/api/cognition/clusters`, `/api/cognition/debt` (studio). A user could scope the health gate but **not** an anchor or drift audit — the capability existed one layer below the surface and was unreachable. All five now take the filter, and the studio gained a labelled **Project Scope** control in the header that drives the health, drift, bug-cluster and technical-debt fetches, seeds from the URL query string (so a scoped link is shareable and reloadable), and renders the report's `scope` back to the user in the CLI's own wording. Cover: `test/health_project_scope_surfaces.test.ts` (7 cases) alongside the original `test/health_project_scope.test.ts` (9 cases).
+- [x] **Wave 24: Per-Project Health Gate Scope**:
+  - `T-321`: The health gate, the bug-clustering scanner and the technical-debt scanner all read the whole store, and the report was then labelled with `entries[0].project` — whichever memory happened to sort first. A number covering every project was presented as one project's health, and there was no way to ask for a single project's health at all. `listForProject()` (`src/store.ts`) is now the single owner of the "filter by project, or take the whole store" decision; the gate, the anchor audit, the drift audit, bug clustering and technical debt all take the same `project` option. The report carries a `scope` object (`filter`, `projects`, `entry_count`, `code_scope`), and its `project` label is the requested filter, the sole project present, or `all` — never a guess. Code-side metrics deliberately stay workspace-wide: source code is shared between projects and cannot be attributed to one. Surfaces: `memory health --project P`, a `project` argument on `muse_health`/`muse_tech_debt`/`muse_bug_clusters`, and `?project=` on the studio's `/api/health` (`test/health_project_scope.test.ts`, 9 cases, commits `60b1aa2`, `27bdc81`).
+- [x] **Wave 23: Contradiction Engine Calibration**:
+  - `T-320`: The contradiction engine compared two memories by testing seven bare regex pairs against their concatenated title and content, so any two memories containing an "enable"-ish and a "disable"-ish word were flagged as contradictory regardless of subject or context. Measured: **8 of 10** realistic non-contradictory pairs were flagged. Flagging flips *both* memories to `conflicted` — 25 health points each (four false positives fail the store-integrity pillar), removal from every future sync broadcast, and still served into agent prompts. Detection now extracts **polarity claims** (`subject`, `relation`, `target`, `scope`) and opposes claims only about the same subject in a compatible scope; re-probe of the same ten pairs: **0 of 10** flagged. New `src/quality/scope.ts` (canonical scope vocabulary, general qualifier extraction, strong/weak scope classification) and `src/quality/polarity.ts` (claim extraction + comparison); scope is now resolved per claim instead of per entry in `src/quality/invalidation.ts`. Dry-run surfaces: `memory conflicts [--dry-run]` (new CLI, read-only preview of what peer-to-peer sync would flag — the one path that flags automatically), `propose({ checkConflict: true, conflictDryRun: true })`, and `detectConflict(store, candidate, { dryRun: true })`. Calibration fixture `test/calibration/contradiction_calibration.json` (21 labelled pairs) gated at precision and recall ≥ 0.9 (`test/contradiction_calibration.test.ts`, `test/cli_conflicts.test.ts`, commits `18bf4f9`, `dc90539`).
 - [x] **Wave 22: Failure Attribution, Negative Memory Mining & Retrieval Compounding**:
   - `T-313`: Scoped failure attribution matching code anchors (`file_path`, `symbol_name`), affected paths, and reproduction commands; session-deduplicated strikes (max 1/session); dual-decay strike counter (2 clean runs or 30 days without failure); constitutional immunity protecting constraints, timeless memories, and salience $\ge 0.95$ from auto-dispute (`test/failure_attribution.test.ts`, commit `41feec4`, `3b270b5`).
   - `T-144`: Trajectory fix-pair mining discovering negative memories (`AVOID: ...`) from *Error ➔ Remediation ➔ No Recurrence* observation sequences with affinity filters excluding non-mutating commands (`ls`, `cat`), dynamic line number normalization, Vibeguard secret scrubbing, and session volume cap of max 3 candidates (`test/fix_pair_mining.test.ts`, commit `cc4b5c5`).
@@ -332,6 +469,9 @@ Removed remaining competitor product names and external comparison links from we
 - **Zero External Daemons**: All operations must execute in-process via SQLite (`memory.db`) or flat files. Never depend on background servers.
 - **Fail Fast**: Throw loudly on secret detection, invalid schema shapes, or unconfirmed supersessions.
 - **Verification Gate**: Before declaring changes complete, `bun test`, `bun run build`, and `tsc --noEmit` must pass.
+- **Workspace Isolation & Sync Boundary**: Never run repo-mutating scaffolders from a parent directory; nested clones are separate scopes. Never sync `.git` or a live SQLite store (`.memory/`).
 
 ## 1. Verified Shipped Reality
+- **2026-10-22**: T-321 per-project health gate scope
+- **2026-10-22**: T-320 contradiction engine calibration
 - **2026-10-20**: Wave 21: LifeOS Distillation, Mutation Safety & Knowledge Routing (T-050, T-177, T-324)

package/.agents/context/decisions.md

@@ -27,3 +27,11 @@
 ### ADR-006: Harness-Independent Hook Installation (Global Standard First)
 - **Decision**: `memory hooks install` resolves its target as explicit `settingsPath` → explicit `--harness` opt-in → global standard (project `.agents/hooks.json` when a `.agents/` directory exists, otherwise `~/.agents/hooks.json`). A harness path (e.g. `~/.claude/settings.json`) is never auto-selected, and `connectAgent` skips harness-specific agents without an explicit `--harness` opt-in.
 - **Rationale** (principal ruling, 2026-10-15): the runtime is vendor-neutral — hooks must not couple memory to one harness or create a harness directory unprompted. A pre-existing harness settings file does not imply consent; only `--harness claude` does.
+
+### ADR-007: Client Scope Descoped from the Memory Store
+- **Decision** (2026-11-09): `scope: client` is not added to `MemoryScope`. Client enclaves remain an explicit encrypted vault (`memory enclave create|list|put|rotate`) outside normal retrieval, recall, and MCP surfaces.
+- **Rationale**: wiring a tenant scope through store/retrieval/sync would make cross-tenant disclosure a single filter bug away. The encrypted vault has its own key lifecycle and is auditable at the CLI boundary. Revisit only with a tenant-aware store substrate and per-client key rotation tests.
+
+### ADR-008: Agent Workspace Tracking Boundary (T-450)
+- **Decision** (2026-11-09): keep `.agents/standards/`, `.agents/skills/`, `.agents/context/`, `.agents/brand/`, `.agents/artifacts/` and `.agents/research/` tracked. Ignore new ephemeral state under `.agents/plans/`, `.agents/archive/`, `.agents/worktree-leases/` and `.agents/scratch/`. Already-tracked files stay tracked for history; the rule applies to new files.
+- **Rationale**: standards, skills, decisions, evidence and audit artifacts are collaborator-facing truth. Plans, leases and scratch are per-agent working state that causes churn and merge noise. Removing already-tracked history would break links in prior tickets, so the boundary moves forward only.

package/.agents/context/deep-research-agentic-memory-evolution.md

@@ -181,7 +181,7 @@ Currently, Muse Memory excels at **Semantic Memory** (facts, rules, ADRs, prefer
 #### B. Multi-Tenant Agency Enclaves (`scope: client`)
 - Tailored for Harsh's agency model with 4 internal divisions: Sol, Jasper, Crew, Nexus.
 - **Isolation Rule**:
-  - `scope: client`: Enclave isolated per client ID (`client_id: "client-acme"`).
+  - `scope: client`: **Descoped** (ADR-007). Enclaves are explicit encrypted vaults; not a store scope.
   - Client A’s confidential database schemas and domain logic are never exposed to Client B.
   - Universal architectural principles (`scope: global`) propagate freely across all client workspaces.
 

package/.agents/context/evidence-ledger.md

@@ -0,0 +1,89 @@
+# Evidence Ledger
+
+## Dashboard
+
+| Category | Count |
+|---|---|
+| DECISION | 1 |
+| CLAIM | 5 |
+| STATUS | 1 |
+
+| Status | Count |
+|---|---|
+| VERIFIED | 4 |
+| QUARANTINED | 1 |
+| ACTIVE | 1 |
+
+**Health score**: 5 / 6 verifiable entries (83%) — one claim is quarantined, not contradicted.
+
+---
+
+### EVD-001 | 2026-11-09 | STATUS
+**Statement**: `memory benchmark` crashed on title-less memories; fixed in the token-budget packing loop.
+**Category**: `STATUS`
+**Confidence**: `[RAW]` — reproduced with `bun bin/musememory.ts benchmark --iterations 5` (TypeError at `item.entry.title.length`), then passed after the null-safe patch.
+**Epistemology**: `[EMPIRICAL]`
+**Evidence**: `src/retrieval/ranking.ts` token estimate; `test/ranking_titleless.test.ts`.
+**Status**: `ACTIVE`
+
+---
+
+### EVD-002 | 2026-11-09 | CLAIM
+**Statement**: On a 34-entry synthetic corpus, retrieval hit the target in the top 3 for 10/10 queries (MRR 0.933, 2.69 ms average, 68% token savings).
+**Category**: `CLAIM`
+**Confidence**: `[RAW]` — scratch benchmark run, output JSON retained in session.
+**Epistemology**: `[EMPIRICAL]`
+**Evidence**: `docs/evidence/2026-11-dogfood.md`; query set and metrics table.
+**Status**: `VERIFIED`
+
+---
+
+### EVD-003 | 2026-11-09 | CLAIM
+**Statement**: On the real 799-memory store, the benchmark reports FTS5 p50 12.81 ms, context assembly p50 15.64 ms, and 99% token compression.
+**Category**: `CLAIM`
+**Confidence**: `[RAW]` — `bun bin/musememory.ts benchmark --iterations 10`.
+**Epistemology**: `[EMPIRICAL]`
+**Evidence**: benchmark table in `docs/evidence/2026-11-dogfood.md`.
+**Status**: `VERIFIED`
+
+---
+
+### EVD-004 | 2026-11-09 | CLAIM
+**Statement**: The engine targets p95 latency: hot cache < 250 µs, FTS5 < 50 ms, context < 50 ms.
+**Category**: `CLAIM`
+**Confidence**: `[RAW]` — 50-iteration benchmark: 176.25 µs, 15.66 ms, 26.69 ms.
+**Epistemology**: `[EMPIRICAL]`
+**Evidence**: `docs/evidence/2026-11-dogfood.md`; benchmark `[TARGET]` line.
+**Status**: `VERIFIED` — the earlier "< 15 ms" claim was a p50 value mislabeled as a bound; it has been corrected.
+
+---
+
+### EVD-005 | 2026-11-09 | CLAIM
+**Statement**: Muse Memory improves coding-agent task success.
+**Category**: `CLAIM`
+**Confidence**: `[INFER]` — plausible from retrieval quality, but no with/without task-outcome study exists.
+**Epistemology**: `[SPECULATIVE]`
+**Evidence**: none yet; retrieval evidence only (EVD-002/003).
+**Status**: `QUARANTINED` — a 2-task pilot (EVD-007) is suggestive, not sufficient.
+
+---
+
+### EVD-006 | 2026-11-09 | DECISION
+**Statement**: Before starting Wave 28, fix the benchmark defect, record evidence, harden the enclave key lifecycle, then cut a release candidate.
+**Category**: `DECISION`
+**Confidence**: `[RAW]` — based on EVD-001 through EVD-005.
+**Epistemology**: `[EMPIRICAL]`
+**Options considered**: (a) continue feature waves; (b) evidence + security + release gate; (c) release immediately without evidence.
+**Trade-offs**: (b) delays new features but converts unverified claims into measurable ones and protects client data.
+**Evidence**: this ledger + `TECHNICAL_DEBT-wave27.md`.
+**Status**: `ACTIVE`
+
+---
+
+### EVD-007 | 2026-11-09 | CLAIM
+**Statement**: In a 2-task pilot, an agent with memory passed 2/2 hidden tests; the same model without memory passed 1/2 (missed the 32-character tag cap).
+**Category**: `CLAIM`
+**Confidence**: `[RAW]` — hidden Bun tests run against both fixture implementations.
+**Epistemology**: `[EMPIRICAL]`
+**Evidence**: `docs/evidence/2026-11-outcome-pilot.md`.
+**Status**: `VERIFIED` — pilot only (n=2, single run).

package/.agents/context/index.md

@@ -13,7 +13,7 @@
 | [`product.md`](./product.md) | Product scope, the live MCP tool matrix, the 7 task profiles, the CLI surface, and the trust gates an agent must assume | Understanding user commands, capabilities, and MCP protocols |
 | [`architecture.md`](./architecture.md) | Dual-Scope storage (`.memory/`), SQLite DB, retrieval scoring formula, and file layout | System design, file locations, state machine, and data flow |
 | [`brand.md`](./brand.md) | Persona profiles (`USER.md` archetypes), voice, and presentation rules | Writing prompts, terminal output copy, and user documentation |
-| [`current.md`](./current.md) | Current shipped state (**v3.0.2**), test coverage, and active constraints | Before starting work on bugs, features, or refactors |
+| [`current.md`](./current.md) | Current shipped state (**v3.1.0**), test coverage, and active constraints | Before starting work on bugs, features, or refactors |
 | [`decisions.md`](./decisions.md) | Locked architectural decisions (zero-daemon, file-backed SQLite, atomic I/O) | Evaluating technical choices or refactors |
 | [`roadmap.md`](./roadmap.md) | Planned improvements, ecosystem expansions, and backlog | Scoping new capabilities |
 | [`ticket-plan.md`](./ticket-plan.md) | Detailed ticket plan for the next work waves | Planning a specific wave of work |

package/.agents/context/roadmap.md

@@ -53,7 +53,7 @@
 - **Git-Aware Branchable Memory Overlays (`memory branch`)**:
   - Branch-isolated working memory overlays that mirror git branch lifecycles (`feature/*`, `fix/*`).
   - Automated `memory branch-merge` command with interactive semantic contradiction resolution when merging feature branches into `main`.
-- **Agency Multi-Tenant Enclaves (`scope: client`)**:
+- **Agency Multi-Tenant Enclaves** (`scope: client` descoped — ADR-007; vaults are explicit):
   - Cryptographically isolated client memory vaults tailored for multi-client agency delivery (Sol, Jasper, Crew, Nexus).
   - Enforces zero cross-client knowledge contamination while allowing global engineering invariants (`scope: global`) to propagate safely.
 - **Zero-Knowledge Memory Attestation**:
@@ -167,7 +167,7 @@
 |---|---|---|
 | **Speed & Observability** | ✅ Shipped (sub-2ms, <50µs hot cache, 527 tests) | R25 (--include-abstract, --scope) |
 | **Hands-free hooks** | ✅ 3/6 events shipped (GAP-2) | R24 (full plugin registry) |
-| **Trust gates** | ✅ Write-time verify + recall-time freshness gate (GAP-1, GAP-3) | R23 (extend to multi-repo) |
+| **Trust gates** | ✅ Write-time verify + recall-time freshness gate (GAP-1, GAP-3) + calibrated contradiction engine (T-320) | R23 (extend to multi-repo) |
 | **Version control** | ⚠️ No branch-aware memory | R23 (memory branch/checkout/blame) |
 | **Compounding** | ✅ 7-state lifecycle + archival ladder | R23 (access-weighted retention, half-life decay) |
 | **Team/SaaS** | ⚠️ Not yet built | R26 (pricing page, Plausible, SaaS prototype) |
@@ -379,3 +379,12 @@ Exploded ticket plan: `.agents/context/ticket-plan.md` (269 tracked tickets, 21
    - `T-153`: Multi-DB topology (shared team store + personal DBs)
    - `T-387`: Client configuration recipes & shared+personal scoping governance
 
+9. **Wave 23: Dashboard Overhaul, Workspace Terrain Maps & High-Speed Agent Hub (v3.1.0 Roadmap)**
+   - `T-451`–`T-453`: Cyber-minimalist left-rail poster navigation shell, dual-scope switcher (`project` | `global` | `mesh`), and sub-10ms UI static route serving
+   - `T-454`–`T-456`: Memory Studio 2.0 with instant Canvas Graph ↔ Filterable List toggle, 60fps kinetic-settling physics loop, and 52-week activity calendar heatmap
+   - `T-457`–`T-459`: Workspace Terrain Generator (`stack.md`, `habits.md`, `ownership.md`, `packages.md`, `repo.md`, `logs.md`) & split-pane markdown reader with scanline reindexing
+   - `T-460`–`T-462`: Universal Agents Hub with live auto-discovery (Antigravity, OpenCode, Claude Code, Cursor, Copilot, Hermes), 1-click MCP wiring, and drop-in prompt snippets
+   - `T-463`–`T-465`: Poster-style About panel (WHAT, THE PIECES live facts table, PRIVACY guarantee), multi-agent quick start guide, and interactive CLI command deck
+   - `T-466`–`T-468`: In-memory SQLite read projection cache, tab visibility sensors for zero idle CPU burn, and headless UI performance benchmark suite
+
+