feat(review): режим --full — закрытие сессии одной командой (v1.2.8)

.claude-plugin/marketplace.json

@@ -10,7 +10,7 @@
     {
       "name": "mnemo",
       "description": "7 memory skills for Claude Code and Codex via Obsidian — vault audit, knowledge search, bounded cross-runtime recall, connection discovery, session handoff, memory routing, review orchestration, and setup",
-      "version": "1.2.7",
+      "version": "1.2.8",
       "source": "./plugins/mnemo",
       "category": "productivity",
       "keywords": [

CHANGELOG.md

@@ -6,6 +6,18 @@ The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 
 ## [Unreleased]
 
+## [1.2.8] - 2026-07-21
+
+### Added
+
+- **`/mn:review --full` — one-command session close-out.** An explicit `--full` flag turns `review` into the whole end-of-session ritual with no per-skill prompt: the flag itself is consent (**not** the implicit autorun removed in v0.16.0 — the user types it). It anchors on the session's **origin** (reconstructs the first request, measures drift: discussed vs wanted vs did), audits the arc, chains **save → session → [focus text] → connect**, then runs a read-only **verify** pass. `health` is excluded (heavy — manual). Folds the four manual "be thorough / did we capture everything / what's left / find hidden links" prompts into built-in phases. Plain `/mn:review` is byte-for-byte unchanged (audit + one interactive offer, never auto-runs).
+- **`references/depth-contract.md` — thoroughness by routing, not volume.** A standing contract `--full` injects into its `save`/`session` steps (and `session` reads as its default bar): business-logic / pains / how-the-user-thinks route into `save`'s typed `principle`/`pain`/`stance` atoms, links to `connect`, unfinished work to the handoff — never a "capture everything" blob. Encodes depth = structure (the v1.2.7 rule) across the whole close-out.
+- **Grounded, idempotent verify guardrails.** The verify pass cites only external facts (git diff, orphans, `session-scan`, a Step-0 snapshot), never self-grades; its non-orphan check is binary and never rewards link count (orphans delegate to `connect`; verify never links); it REPORTS a missing prod/e2e verification as an unchecked gap (memory-not-CI — it never runs QA); and a re-run on an unchanged session prints "already in order" and stops.
+
+### Changed
+
+- **`session` is thorough-by-routing and self-checks its own note.** Step 1 carries a standing depth-routing default (pointing at the depth-contract) so even a standalone `/mn:session` routes atom-worthy material to `save` instead of swelling the narrative; Step 7 adds an own-note self-check (dup / MOC / orphan / delegation). The cross-skill palace audit stays in `review --full`. `review`'s description gains the `--full` triggers; `evals/trigger-eval.json` adds P9/P10 (positives) + N8 (near-miss); `test-skill-write-contracts.py` pins the new `--full` invariants (flag = consent not autorun, chain order, grounded-binary-never-links verify, depth-contract routing).
+
 ## [1.2.7] - 2026-07-21
 
 ### Added
@@ -911,7 +923,8 @@ Frontmatter now includes `session_id: {CLAUDE_SESSION_ID}` — disambiguates sam
 - `config.example.json`
 - MIT License
 
-[Unreleased]: https://github.com/jojoprison/mnemo/compare/v1.2.7...HEAD
+[Unreleased]: https://github.com/jojoprison/mnemo/compare/v1.2.8...HEAD
+[1.2.8]: https://github.com/jojoprison/mnemo/compare/v1.2.7...v1.2.8
 [1.2.7]: https://github.com/jojoprison/mnemo/compare/v1.2.6...v1.2.7
 [1.2.6]: https://github.com/jojoprison/mnemo/compare/v1.2.5...v1.2.6
 [1.2.5]: https://github.com/jojoprison/mnemo/compare/v1.2.4...v1.2.5

docs/design-decisions.md

@@ -44,6 +44,21 @@ mnemo saves two different kinds of thing, and they must land in different places
 - **Auto-route on explicit save, confirmed in the orchestrator.** A direct `/mn:save` of a rule routes automatically (the user explicitly asked to save). `/mn:review` never writes unattended — since v0.16.0 *every* skill it wants to run goes through one Step 8 confirmation, and an actionable rule additionally gets its own explicit line item there, because routing it creates/edits committed project files.
 - **Codex caveat acknowledged, not solved here.** Codex reads only `AGENTS.md` (32 KiB, silent truncate), not `.claude/rules/`. mnemo routes to `.claude/rules/` and flags the AGENTS.md build-step as the user's responsibility — it does not own that assemble step.
 
+## One-command close-out: `review --full` (v1.2.8)
+
+`/mn:review --full` collapses the end-of-session ritual — audit, save, session, connect, verify — into a single explicit command, so the user never pastes a "capture everything, check everything, keep it like a clean palace" wall of prompt. It is a **flag on `review`**, not an eighth skill (the 7-skill canon holds).
+
+**Deliberate boundaries** (each is load-bearing — removing it reintroduces a rejected behavior):
+
+- **Flag = consent, not implicit autorun.** The explicit `--full` the user typed *is* the confirmation, so the chain runs without a per-skill `y`. This does not revive the autorun removed in v0.16.0: that was *plain* `/mn:review` firing skills unasked. Default `/mn:review` still audits + offers + never auto-runs — only the typed flag chains.
+- **Verify grounds externally, never self-grades.** The Step 9 verify pass anchors every check to a fact — git diff, orphans, `session-scan`, the Step-0 snapshot — never the agent's own assertion. A same-agent self-audit that trusts itself rubber-stamps (eval-integrity).
+- **Verify never links; over-linking is defused.** The non-orphan check is binary (connected or not) and never rewards link *count*; an orphan is delegated to `connect`, and verify adds no link itself. So "score the palace green" cannot incentivize link-spam from a "connect everything" instruction.
+- **Depth-contract is routing, not volume.** `--full` injects `references/depth-contract.md` as guidance: thorough means the right material in the right home (business-logic / pains / mental-model → `save`'s typed atoms, **not** the session narrative), never a bigger note. "Capture everything, super-detailed" is the blob anti-pattern rejected in v1.2.7; it is explicitly dropped.
+- **memory-not-CI.** For "did we verify on prod / run e2e / really done?", `--full` REPORTS the absence of test/deploy evidence as an unchecked gap; it never runs tests, hits prod, or fires a trigger. That execution lives in the harness (`finish-the-work` / `loop-gate`), outside a memory plugin's mandate.
+- **`health` stays manual.** Heavy; recommended in the report, never auto-chained.
+- **`session` stays pure narrative.** Its own Step 7 self-check verifies only *its* note (dup / MOC / orphan / delegation); the cross-skill palace audit is `review --full`'s Step 9, not session's — the same boundary that keeps rule-routing out of session (v0.15.0).
+- **Idempotent.** A second `--full` on an unchanged session prints "already in order" and stops — it never re-parks or re-links.
+
 ## Proactive nudges via hooks (v1.1.1)
 
 Descriptions get an agent to *consider* mnemo, but Opus 4.8 / Fable 5 under-trigger skills — a good description raises the odds, it doesn't guarantee the call. v1.1.1 adds a deterministic **delivery** layer via hooks, with one honest caveat baked into the design: **deterministic delivery ≠ deterministic effect** — the hook always fires, but the model still decides whether to act. A prose nudge is exactly the "marginal rule ≈ 0" pattern, so it's kept short, factual (not an order), and gated.

docs/review.md

@@ -2,7 +2,7 @@
 
 ## Overview
 
-The only command you need at session end. Analyzes everything, then presents one prioritized offer — unsaved decisions (`/mn:save`), session notes (`/mn:session`), and the remaining skills. **Nothing runs without your confirmation.** One command replaces manually running save + session + connect + health — you approve the batch once.
+The only command you need at session end. Analyzes everything, then presents one prioritized offer — unsaved decisions (`/mn:save`), session notes (`/mn:session`), and the remaining skills. **Nothing runs without your confirmation.** One command replaces manually running save + session + connect + health — you approve the batch once. Add `--full` and even that one approval goes away: the flag itself is consent (see [`--full` below](#--full--one-command-close-out-v128)).
 
 ## Usage
 
@@ -115,6 +115,21 @@ Recommended:
 Run any? (1,2,3,4 / A=all / N=skip)
 ```
 
+## `--full` — one-command close-out (v1.2.8)
+
+```
+/mn:review --full
+```
+
+The explicit flag **is** the consent, so the whole close-out runs with no per-skill prompt. It:
+
+1. **Anchors on the session's origin** — reconstructs the first request and measures drift (discussed vs wanted vs did), not just an end-state inventory.
+2. **Audits** the whole arc (Done / Missed / Hanging Threads / score) — this produces the recommendation list.
+3. **Executes the chain** in order — `save → session → [any focus text you passed after --full] → connect` — injecting the depth-contract so business-logic / pains / how-you-think route into typed `save` atoms, not the session narrative. `health` is excluded (heavy — run it by hand).
+4. **Verifies (read-only)** — checks the freshly-parked notes for typed-slot quality, atomicity, and connectedness (binary orphan check, delegated to `connect`, never self-linked); REPORTS any missing prod/e2e verification as a gap (mnemo never runs QA); and is idempotent — a second `--full` on an unchanged session prints "🏛 already in order" and stops.
+
+Plain `/mn:review` (no flag) is unchanged: audit + one interactive offer, never auto-runs (v0.16.0). `--full` does **not** revive that killed implicit autorun — the user typed the flag.
+
 ## Important Notes
 
 - **Always thorough** — no quick mode, full analysis every time

docs/session.md

@@ -20,6 +20,7 @@ No arguments. Summarizes the current conversation automatically.
 4. Verifies the note is linked in the hub reached through `taxonomy_roles.moc`
 5. Updates `Meta — Session Handoff` with pending items
 6. Checks for orphans after creation
+7. Self-checks its own note (duplicate / MOC link / orphan / atom-delegation) before confirming
 
 ## Example Output
 
@@ -62,6 +63,8 @@ When the next session starts, it reads `Meta — Session Handoff`:
 - **MOC verification** — automatically adds to MOC if missing
 - **Branch field optional** — research sessions don't have branches
 - **Ghost notes generously** — wraps projects, technologies, people in `[[wikilinks]]`
+- **Thorough by routing, not volume** — the note stays a narrative; decisions, business-logic, pains, and how-you-think route to `save`'s typed atoms (see the depth-contract), links to `connect`, unfinished work to handoff — never a "capture everything" blob
+- **Own-note self-check** — Step 7 verifies only this note's own artifact; the cross-skill palace audit is `review --full`'s job, not session's
 
 ## Related Skills
 

evals/trigger-eval.json

@@ -1,5 +1,5 @@
 {
-  "about": "Trigger-eval set for mnemo skill descriptions. 8 proactive-positive prompts (a specific skill SHOULD auto-invoke) + 7 near-miss negatives (share keywords but must NOT trigger). Re-run after editing any canonical description to catch trigger regressions. Baseline 2026-07-05 (v1.1.x): 12/12 at the simulated-routing level. 2026-07-21 (v1.2.7): +P7/P8 for the save W4 principle·pain·stance layer, +N7 near-miss (a code-comment 'принцип', not a personal-memory save). Format is compatible with skill-creator's run_loop.py (query + should_trigger); expect_skill/why are extra context.",
+  "about": "Trigger-eval set for mnemo skill descriptions. 10 proactive-positive prompts (a specific skill SHOULD auto-invoke) + 8 near-miss negatives (share keywords but must NOT trigger). Re-run after editing any canonical description to catch trigger regressions. Baseline 2026-07-05 (v1.1.x): 12/12 at the simulated-routing level. 2026-07-21 (v1.2.7): +P7/P8 for the save W4 principle·pain·stance layer, +N7 near-miss (a code-comment 'принцип', not a personal-memory save). 2026-07-21 (v1.2.8): +P9/P10 for the review --full one-command close-out, +N8 near-miss ('полный прогон тестов' — 'полный' overlaps --full but is CI, not review). Format is compatible with skill-creator's run_loop.py (query + should_trigger); expect_skill/why are extra context.",
   "evals": [
     { "id": "P1", "query": "Сейчас буду чинить баг в auth — токен опять отваливается на рефреше, вроде уже сталкивались с похожим", "should_trigger": true, "expect_skill": "ask", "why": "recurring/previously-seen bug → recall before re-fixing" },
     { "id": "P2", "query": "Разобрался почему CI падал: race condition в setup-фикстуре, добавил await перед teardown. Наконец-то", "should_trigger": true, "expect_skill": "save", "why": "solved a tricky bug → proactive save" },
@@ -9,12 +9,15 @@
     { "id": "P6", "query": "найди связи для только что созданной заметки про OAuth", "should_trigger": true, "expect_skill": "connect", "why": "explicit connect + new note" },
     { "id": "P7", "query": "запомни как я обычно принимаю решения — сначала ambitious-вариант с честными trade-offs, потом safe", "should_trigger": true, "expect_skill": "save", "why": "the user's decision stance / mental model → W4 principle·stance save" },
     { "id": "P8", "query": "вот моя боль: каждую сессию вставляю огромную портянку промпта чтобы получить нормальный save", "should_trigger": true, "expect_skill": "save", "why": "a user pain / business-logic → W4 pain save" },
+    { "id": "P9", "query": "закрой сессию начисто одной командой — прогони весь цикл закрытия и проверь что всё уложено", "should_trigger": true, "expect_skill": "review", "why": "one-command end-of-session close-out → review (--full)" },
+    { "id": "P10", "query": "review --full", "should_trigger": true, "expect_skill": "review", "why": "explicit --full flag → review one-command pass" },
     { "id": "N1", "query": "найди функцию parse_order в кодовой базе и покажи где она вызывается", "should_trigger": false, "why": "grep of CODE, not vault recall — near-miss ask" },
     { "id": "N2", "query": "поищи в интернете как работает OAuth PKCE flow, дай ссылки", "should_trigger": false, "why": "web search, not vault — 'поищи' overlaps, meaning differs" },
     { "id": "N3", "query": "сохрани этот файл и закоммить", "should_trigger": false, "why": "'сохрани' is about a FILE + VCS, not memory — near-miss save" },
     { "id": "N4", "query": "свяжись с бэкенд-командой насчёт API-контракта", "should_trigger": false, "why": "'свяжись' = communicate, not 'link notes' — near-miss connect" },
     { "id": "N5", "query": "проверь здоровье сервиса — дёрни /health эндпоинт прода", "should_trigger": false, "why": "service health, not vault health — near-miss health" },
     { "id": "N6", "query": "почини опечатку в переменной userNmae → userName", "should_trigger": false, "why": "trivial fix, not a recurring bug — ask must NOT fire" },
-    { "id": "N7", "query": "запиши в комментарии к функции принцип работы алгоритма сортировки", "should_trigger": false, "why": "'принцип'/'запиши' about a CODE comment, not the user's personal mental model — near-miss save (W4)" }
+    { "id": "N7", "query": "запиши в комментарии к функции принцип работы алгоритма сортировки", "should_trigger": false, "why": "'принцип'/'запиши' about a CODE comment, not the user's personal mental model — near-miss save (W4)" },
+    { "id": "N8", "query": "запусти полный прогон тестов перед мержем", "should_trigger": false, "why": "'полный' overlaps review --full but this is a CI test run, not a session close-out — near-miss review" }
   ]
 }

plugins/mnemo/.claude-plugin/plugin.json

@@ -1,7 +1,7 @@
 {
   "name": "mn",
   "displayName": "mnemo",
-  "version": "1.2.7",
+  "version": "1.2.8",
   "description": "Persistent memory layer for Claude Code and Codex via Obsidian — 7 skills for vault health, knowledge search, verified cross-runtime project recall, connection discovery, session continuity, memory routing, skill-aware session review, and interactive setup. Bundled JSON-stdin atomic vault writer, argv-safe indexed reads, dual-runtime helpers, hybrid model routing, parallel CLI calls, SessionStart prewarm, incremental JSONL parsing, shared references, and semantic taxonomy roles.",
   "author": {
     "name": "jojoprison",

plugins/mnemo/.codex-plugin/plugin.json

@@ -1,6 +1,6 @@
 {
   "name": "mnemo",
-  "version": "1.2.7",
+  "version": "1.2.8",
   "description": "Persistent memory layer for Codex and Claude Code via Obsidian: vault health, knowledge search, verified cross-runtime project recall, link discovery, session notes, memory routing, review orchestration, and setup.",
   "author": {
     "name": "jojoprison",