claudenv 1.2.5 → 1.3.1

package/README.md

@@ -1,6 +1,10 @@
 # claudenv
 
-Set up [Claude Code](https://docs.anthropic.com/en/docs/claude-code) in any project with one command. claudenv analyzes your codebase and generates everything Claude needs to work effectively — documentation, rules, hooks, MCP servers, and slash commands.
+[![npm](https://img.shields.io/npm/v/claudenv.svg)](https://www.npmjs.com/package/claudenv)
+
+Set up [Claude Code](https://docs.anthropic.com/en/docs/claude-code) in any project with one command. claudenv analyzes your codebase and generates everything Claude needs to work effectively — documentation, rules, hooks, MCP servers, slash commands, **plus cross-session memory and vibe-decisions logging (1.3.0).**
+
+> **New in 1.3.0** — split memory layout (`~/.claudenv/memories/` global + `.claude/memories/` project), `vibe-decisions` skill (auto-log in loop, pause-and-ask interactive), new CLI: `memory`, `decisions`, `canon`, `doctor`. Companion Python package `claudenv-memory` on PyPI (alpha). See [CHANGELOG.md](./CHANGELOG.md).
 
 ## Quick Start
 
@@ -17,8 +21,8 @@ Claude reads your code, asks a few questions, and generates:
 - **CLAUDE.md** — project overview, architecture, key commands
 - **Rules** — coding style, testing patterns, workflow guidelines (`.claude/rules/`)
 - **MCP servers** — auto-detected from your stack, configured in `.mcp.json`
-- **Slash commands** — `/init-docs`, `/update-docs`, `/validate-docs`, `/setup-mcp`, `/improve`
-- **Hooks** — validation on tool use, audit logging (`.claude/settings.json`)
+- **Slash commands** — `/init-docs`, `/update-docs`, `/validate-docs`, `/setup-mcp`, `/improve`, **`/deeper`**, **`/why`**, **`/decisions`**, **`/canon`**, **`/just-code`** (1.3.0)
+- **Hooks** — validation on tool use, audit logging, decisions-logger (PostToolUse Write), regen-index (SessionEnd) — `.claude/settings.json`
 
 Everything is committed to your repo. Team members get the same Claude experience.
 
@@ -151,6 +155,87 @@ claudenv loop --profile moderate --goal "add types"    # sonnet, deny-list guard
 claudenv loop --profile ci --model sonnet              # ci profile but with sonnet
 ```
 
+## Memory & vibe-decisions (1.3.0)
+
+Persistent memory across sessions + brief overview before non-trivial technical choices. Auto-log inside `claudenv loop` (no pauses — `autonomy=law`), pause-and-ask in interactive Claude Code chats.
+
+### Layout (split)
+
+```
+~/.claudenv/memories/              # global, cross-project
+├── INDEX.md                       # briefing — auto-regenerated
+├── decisions/                     # universal tech decisions
+├── canon/index.yaml               # personal canon of references
+└── user/preferences.md            # cross-project preferences
+
+<project>/.claude/memories/        # project-scoped, committed
+├── project.md                     # stack, conventions, internal urls
+└── decisions/                     # project-only decisions
+```
+
+### Commands
+
+```bash
+# Memory
+claudenv memory init                     # initialise ~/.claudenv/memories/
+claudenv memory index                    # regenerate INDEX.md
+claudenv memory show <path>              # print a memory file
+claudenv memory edit <path>              # open in $EDITOR
+
+# Decisions
+claudenv decisions list [--scope all|global|project] [--limit N]
+claudenv decisions show <id-or-slug>
+claudenv decisions search <query>
+claudenv decisions archive <id>
+
+# Canon
+claudenv canon add <topic> <url> --why "<reason>" [--title <t>] [--author <a>]
+claudenv canon list [<topic>]
+claudenv canon search <query>
+claudenv canon prune --months 6          # find stale entries
+
+# Health
+claudenv doctor                          # OK/WARN/FAIL check
+```
+
+### Slash-commands (in Claude Code session)
+
+| Command | Purpose |
+|---|---|
+| `/deeper` | Develop the most recent decision into a full deep dive (concept → variants → trade-offs → canon) |
+| `/why <X>` | Explain technology X without logging a decision |
+| `/decisions` | List/search logged decisions via the CLI |
+| `/canon` | Browse the personal canon |
+| `/just-code` | Suppress vibe-decisions overview for the next response only |
+
+### How auto-log works in `claudenv loop`
+
+`claudenv loop` appends a system-prompt fragment that switches the `vibe-decisions` skill into **auto-log mode** — it picks the approach, writes the decision file (`/memories/decisions/<date>-<slug>.md`), and continues coding without pausing. Outside loop (interactive Claude Code) the skill defaults to **pause-and-ask** mode.
+
+After each iteration, `claudenv loop` regenerates `INDEX.md` so the next iteration's briefing reflects what was just decided.
+
+### Python companion: `claudenv-memory`
+
+Building your own agent on Claude Agent SDK and want the same memory layout?
+
+```bash
+pip install claudenv-memory
+```
+
+```python
+from claudenv_memory import LocalFileMemoryBackend
+from claude_agent_sdk import ClaudeAgentOptions
+
+backend = LocalFileMemoryBackend()  # ~/.claudenv/memories + ./.claude/memories
+options = ClaudeAgentOptions(
+    model="claude-opus-4-7",
+    extra_headers={"anthropic-beta": "context-management-2025-06-27"},
+    tools=[backend.as_tool()],
+)
+```
+
+Status: 0.1.x alpha — first production consumer is claudenv 2.0. See `python/README.md`.
+
 ## MCP Server Setup
 
 `/claudenv` auto-detects your tech stack and recommends MCP servers from the [official registry](https://registry.modelcontextprotocol.io). You can also run `/setup-mcp` independently.
@@ -221,6 +306,11 @@ claudenv autonomy [-p <profile>]      Configure autonomy profiles
 claudenv init [dir] [-y]              Legacy: static analysis (no AI)
 claudenv generate [-d <dir>]          Templates only, no scaffold
 claudenv validate [-d <dir>]          Check documentation completeness
+
+claudenv memory init|index|show|edit  Manage ~/.claudenv/memories/ (1.3.0+)
+claudenv decisions list|show|search   Logged vibe-decisions (1.3.0+)
+claudenv canon add|list|search|prune  Personal canon (1.3.0+)
+claudenv doctor                       Health check (1.3.0+)
 ```
 
 ## Run Without Installing

package/bin/cli.js

@@ -288,6 +288,300 @@ program
   .option('--dry-run', 'Preview without writing')
   .action(runAutonomy);
 
+// --- hook (internal entry point for Claude Code hooks) ---
+program
+  .command('hook')
+  .description('Internal: dispatch a Claude Code hook by name (decisions-logger, regen-index)')
+  .argument('<name>', 'Hook name')
+  .action(async (name) => {
+    const { dispatch } = await import('../src/hooks/dispatcher.js');
+    await dispatch(name);
+  });
+
+// =============================================
+// 1.3.0: memory / decisions / canon / doctor
+// =============================================
+
+// --- memory ---
+const memoryCmd = program.command('memory').description('Manage the global ~/.claudenv/memories/ layout');
+
+memoryCmd
+  .command('init')
+  .description('Create the ~/.claudenv/memories/ structure (idempotent)')
+  .action(async () => {
+    const { memoryInit } = await import('../src/memory.js');
+    const { created, skipped } = await memoryInit();
+    for (const p of created) console.log(`  + ${p}`);
+    for (const p of skipped) console.log(`  ~ ${p}`);
+    console.log(`\n  ${created.length} created, ${skipped.length} already present.`);
+  });
+
+memoryCmd
+  .command('index')
+  .description('Regenerate ~/.claudenv/memories/INDEX.md from current decisions and prefs')
+  .action(async () => {
+    const { memoryIndex } = await import('../src/memory.js');
+    const result = await memoryIndex();
+    console.log(`  INDEX.md regenerated → ${result.indexPath}`);
+    console.log(`  ${result.recentCount} recent of ${result.decisionCount} total decisions`);
+  });
+
+memoryCmd
+  .command('show')
+  .argument('<path>', 'Path relative to ~/.claudenv/memories/')
+  .description('Print a memory file to stdout')
+  .action(async (path) => {
+    const { memoryShow } = await import('../src/memory.js');
+    try {
+      process.stdout.write(await memoryShow(path));
+    } catch (err) {
+      console.error(err.message);
+      process.exit(2);
+    }
+  });
+
+memoryCmd
+  .command('edit')
+  .argument('<path>', 'Path relative to ~/.claudenv/memories/')
+  .description('Open a memory file in $EDITOR (vi by default)')
+  .action(async (path) => {
+    const { memoryEdit } = await import('../src/memory.js');
+    const result = await memoryEdit(path);
+    process.exit(result.exitCode);
+  });
+
+// --- decisions ---
+const decisionsCmd = program.command('decisions').description('List, show, or search logged vibe-decisions');
+
+decisionsCmd
+  .command('list')
+  .description('List recent decisions (newest first)')
+  .option('--scope <s>', 'global | project | all', 'all')
+  .option('--limit <n>', 'Max entries', '10')
+  .action(async (opts) => {
+    const { listDecisions, formatDecisionList } = await import('../src/decisions.js');
+    const limit = parseInt(opts.limit, 10) || 10;
+    const all = await listDecisions({ scope: opts.scope });
+    console.log(formatDecisionList(all.slice(0, limit)));
+  });
+
+decisionsCmd
+  .command('show')
+  .argument('<id>', 'Slug or substring of the decision')
+  .description('Show full details of one decision')
+  .action(async (id) => {
+    const { showDecision, formatDecisionDetail } = await import('../src/decisions.js');
+    try {
+      const d = await showDecision(id);
+      console.log(formatDecisionDetail(d));
+      console.log('\n---\n');
+      console.log(d.text);
+    } catch (err) {
+      console.error(err.message);
+      process.exit(err.notFound ? 1 : 2);
+    }
+  });
+
+decisionsCmd
+  .command('search')
+  .argument('<query>', 'Substring to search topic/reason/chose')
+  .description('Search decisions')
+  .action(async (query) => {
+    const { searchDecisions, formatDecisionList } = await import('../src/decisions.js');
+    const hits = await searchDecisions(query);
+    console.log(formatDecisionList(hits));
+  });
+
+decisionsCmd
+  .command('archive')
+  .argument('<id>', 'Slug to archive')
+  .description('Move a decision into <scope-dir>/archive/')
+  .action(async (id) => {
+    const { archiveDecision } = await import('../src/decisions.js');
+    try {
+      const { from, to } = await archiveDecision(id);
+      console.log(`  archived: ${from} → ${to}`);
+    } catch (err) {
+      console.error(err.message);
+      process.exit(2);
+    }
+  });
+
+// --- canon ---
+const canonCmd = program.command('canon').description('Personal canon of references (~/.claudenv/memories/canon/index.yaml)');
+
+canonCmd
+  .command('add')
+  .argument('<topic>', 'Topic slug')
+  .argument('<url>', 'URL to add')
+  .option('--why <reason>', 'Why this reference is in the canon')
+  .option('--title <title>', 'Title')
+  .option('--author <author>', 'Author or venue')
+  .action(async (topic, url, opts) => {
+    const { canonAdd } = await import('../src/canon.js');
+    try {
+      const res = await canonAdd({
+        topic,
+        url,
+        why: opts.why,
+        title: opts.title,
+        author: opts.author,
+      });
+      if (res.added) console.log(`  + ${topic}: ${res.entry.title || res.entry.url}`);
+      else console.log(`  ~ duplicate url in ${topic} — skipped`);
+    } catch (err) {
+      console.error(err.message);
+      process.exit(2);
+    }
+  });
+
+canonCmd
+  .command('list')
+  .argument('[topic]', 'Optional topic filter')
+  .action(async (topic) => {
+    const { canonList, formatCanon } = await import('../src/canon.js');
+    const data = await canonList(topic);
+    console.log(formatCanon(data));
+  });
+
+canonCmd
+  .command('search')
+  .argument('<query>', 'Substring search')
+  .action(async (query) => {
+    const { canonSearch, formatCanon } = await import('../src/canon.js');
+    const data = await canonSearch(query);
+    console.log(formatCanon(data));
+  });
+
+canonCmd
+  .command('prune')
+  .option('--months <n>', 'Age threshold in months', '6')
+  .action(async (opts) => {
+    const { canonPrune } = await import('../src/canon.js');
+    const stale = await canonPrune(parseInt(opts.months, 10) || 6);
+    if (stale.length === 0) {
+      console.log('  No stale entries.');
+      return;
+    }
+    for (const { topic, entry, reason } of stale) {
+      const why = reason || `added ${entry.added}`;
+      console.log(`  ${topic}: ${entry.url} (${why})`);
+    }
+  });
+
+// =============================================
+// Workspaces (isolated per-company/context memory)
+// =============================================
+const wsCmd = program.command('workspace').description('Isolated memory spaces (~/.claudenv/workspaces/)');
+
+wsCmd
+  .command('add')
+  .argument('<id>', 'Workspace id (slug: a-z0-9-_)')
+  .option('--name <name>', 'Human-readable name')
+  .option('--desc <description>', 'Description')
+  .option('--path <path...>', 'Project path(s) bound to this workspace')
+  .action(async (id, opts) => {
+    const { createWorkspace } = await import('../src/workspaces.js');
+    try {
+      await createWorkspace(id, { name: opts.name, description: opts.desc, paths: opts.path });
+      console.log(`  + workspace "${id}" created`);
+    } catch (err) {
+      console.error(err.message);
+      process.exit(2);
+    }
+  });
+
+wsCmd
+  .command('list')
+  .action(async () => {
+    const { listWorkspaces } = await import('../src/workspaces.js');
+    const list = await listWorkspaces();
+    if (list.length === 0) {
+      console.log('  No workspaces yet — create one with `claudenv workspace add <id>`');
+      return;
+    }
+    for (const w of list) {
+      console.log(`  ${w.active ? '*' : ' '} ${w.id}${w.name && w.name !== w.id ? ` — ${w.name}` : ''}`);
+    }
+  });
+
+wsCmd
+  .command('use')
+  .argument('<id>', 'Workspace id to activate')
+  .action(async (id) => {
+    const { useWorkspace } = await import('../src/workspaces.js');
+    try {
+      await useWorkspace(id);
+      console.log(`  active workspace → ${id}`);
+    } catch (err) {
+      console.error(err.message);
+      process.exit(2);
+    }
+  });
+
+wsCmd
+  .command('show')
+  .action(async () => {
+    const { showActive } = await import('../src/workspaces.js');
+    const w = await showActive();
+    if (!w) {
+      console.log('  No active workspace (set CLAUDENV_WORKSPACE or run `claudenv workspace use <id>`)');
+      return;
+    }
+    console.log(`  active: ${w.id} (${w.source})`);
+    if (w.name && w.name !== w.id) console.log(`  name:   ${w.name}`);
+    if (w.description) console.log(`  desc:   ${w.description}`);
+    if (w.paths && w.paths.length) console.log(`  paths:  ${w.paths.join(', ')}`);
+  });
+
+// =============================================
+// Sources (connectors in the active workspace)
+// =============================================
+const srcCmd = program.command('source').description('Data-source connectors in the active workspace');
+
+srcCmd
+  .command('list')
+  .action(async () => {
+    const { listConnectors } = await import('../src/sources.js');
+    const { workspace, connectors } = await listConnectors();
+    if (!workspace) {
+      console.log('  No active workspace — set one first (`claudenv workspace use <id>`)');
+      return;
+    }
+    if (connectors.length === 0) {
+      console.log(`  [${workspace}] no connectors yet — add one via /add-source in Claude Code`);
+      return;
+    }
+    console.log(`  [${workspace}]`);
+    for (const c of connectors) {
+      console.log(`    ${c.name}  (${c.type}, ${c.status})${c.host ? `  ${c.host}` : ''}`);
+    }
+  });
+
+srcCmd
+  .command('show')
+  .argument('<name>', 'Connector name')
+  .action(async (name) => {
+    const { showConnector } = await import('../src/sources.js');
+    const text = await showConnector(name);
+    if (!text) {
+      console.error(`  Connector "${name}" not found in active workspace`);
+      process.exit(2);
+    }
+    console.log(text);
+  });
+
+// --- doctor ---
+program
+  .command('doctor')
+  .description('Health-check the claudenv setup')
+  .action(async () => {
+    const { runDoctor } = await import('../src/doctor.js');
+    const { lines, hasFail } = await runDoctor();
+    for (const l of lines) console.log(l);
+    process.exit(hasFail ? 1 : 0);
+  });
+
 // =============================================
 // Install / Uninstall
 // =============================================

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "claudenv",
-  "version": "1.2.5",
+  "version": "1.3.1",
   "description": "One command to integrate Claude Code into any project — installs /claudenv command for AI-powered documentation generation",
   "type": "module",
   "bin": {

package/scaffold/.claude/memories/README.md

@@ -0,0 +1,44 @@
+# Project memories
+
+This directory holds **project-scoped** memory for `vibe-decisions` and related skills. Cross-project memory lives in `~/.claudenv/memories/`.
+
+## Layout
+
+```
+.claude/memories/
+├── project.md           # стек, конвенции, важные urls — заполняешь вручную
+├── decisions/           # project-specific tech decisions (committed to repo)
+│   └── 2026-05-27-auth-jwt.md
+└── README.md            # this file
+```
+
+## When goes here vs `~/.claudenv/memories/`
+
+- **Здесь (project)**: выбор специфичен этому проекту — внутренний сервис, формат API, deployment target, project-only convention
+- **В `~/.claudenv/memories/decisions/` (global)**: универсальный tech-выбор — какая БД, какой algo, который применим в любом проекте
+
+Skill `vibe-decisions` решает scope автоматически по полю `scope: global|project` в frontmatter. По умолчанию `global` если непонятно.
+
+## project.md шаблон
+
+Создай этот файл вручную (claudenv не генерит автоматически):
+
+```markdown
+# <project-name>
+
+- **Stack:** Python 3.12, FastAPI, Postgres, Redis
+- **Tests:** `pytest -m "not slow"` локально; CI запускает всё
+- **Package manager:** uv (NOT pip, NOT poetry)
+- **Линтеры:** ruff + mypy strict
+- **Deploy:** GitHub Actions → ArgoCD → k8s
+- **Внутренние ADR:** `docs/adr/`
+- **Слаг для decisions/:** `<project-slug>`
+```
+
+`vibe-decisions` читает project.md перед каждым нетривиальным выбором.
+
+## Commit policy
+
+- `project.md` — committed
+- `decisions/*.md` — committed (это shared с командой)
+- Никаких secrets (production endpoints, API keys, internal hostnames) — для них используй `~/.claudenv/memories/` с глобальным scope (выйдет за пределы repo)

package/scaffold/global/.claude/commands/add-source.md

@@ -0,0 +1,31 @@
+---
+description: Добавить коннектор к источнику данных (SQL/REST/вики/Redash) с изоляцией по workspace
+allowed-tools: Bash, Read, Write, Edit, Skill
+argument-hint: [<описание источника>]
+---
+
+# /add-source - подключить источник данных
+
+Запускает skill `source-connector`: опрашивает тебя, генерирует рабочий коннектор,
+кладёт секреты в `.env.local`, кэширует знание о коннекторе (параметры + provenance)
+в память активного workspace.
+
+## Routing
+
+1. Активируй skill `source-connector`.
+2. Если в `$ARGUMENTS` есть описание источника (тип/хост/система) - передай его как
+   стартовый контекст, остальное доспроси.
+3. Если аргументов нет - начни с вопроса, к какому источнику подключаемся.
+
+## Перед стартом
+
+- Определи активный workspace (env `CLAUDENV_WORKSPACE` -> файл
+  `~/.claudenv/active-workspace`). Нет активного - предложи выбрать/создать,
+  не пиши в глобальную память.
+- Напомни инвариант: секреты идут только в `.env.local` (gitignored), в память -
+  только имена переменных.
+
+## Когда НЕ создавать коннектор
+
+Если для источника есть готовый MCP-сервер - предложи путь через `/setup-mcp`
+(`.mcp.json`), коннектор нужен только для источников без MCP.

package/scaffold/global/.claude/commands/canon.md

@@ -0,0 +1,33 @@
+---
+description: Показать или добавить ссылки в личный канон (~/.claudenv/memories/canon/index.yaml)
+allowed-tools: Bash(claudenv:*), Read
+argument-hint: [list [<topic>]|search <query>|add <topic> <url> --why "<reason>"]
+---
+
+# /canon — Личный канон ссылок
+
+Wrapper над `claudenv canon` CLI.
+
+## Routing
+
+Разбери `$ARGUMENTS`:
+
+- пусто или `list` → `claudenv canon list`
+- `list <topic>` → `claudenv canon list <topic>`
+- `search <query>` → `claudenv canon search "<query>"`
+- `add <topic> <url> --why "<reason>"` → `claudenv canon add <topic> <url> --why "<reason>"`
+- иначе если выглядит как topic slug → `claudenv canon list <slug>`
+- иначе → `claudenv canon search "$ARGUMENTS"`
+
+## Output
+
+Печатай stdout как есть. На non-zero exit code покажи stderr и подскажи usage.
+
+## Когда канон пуст
+
+Если `claudenv canon list` ничего не вернул:
+
+```
+Канон пока пуст. Добавь первые записи когда будешь делать /deeper —
+после deep dive предложу `claudenv canon add <topic> <url> --why "..."`.
+```

package/scaffold/global/.claude/commands/decisions.md

@@ -0,0 +1,29 @@
+---
+description: Показать или найти принятые vibe-decisions (последние N, по slug, или по поисковому запросу)
+allowed-tools: Bash(claudenv:*), Read
+argument-hint: [list|show <id>|search <query>] [--limit N] [--scope global|project|all]
+---
+
+# /decisions — Список и поиск принятых решений
+
+Wrapper над `claudenv decisions` CLI. Запускается из Bash, результат показывается пользователю как есть.
+
+## Routing
+
+Разбери `$ARGUMENTS`:
+
+- пусто или `list` → `claudenv decisions list --limit 10`
+- `list --limit N` → `claudenv decisions list --limit N`
+- `list --scope <s>` → `claudenv decisions list --scope <s>`
+- `show <id>` → `claudenv decisions show <id>`
+- `search <query>` → `claudenv decisions search "<query>"`
+- иначе если выглядит как id или slug → `claudenv decisions show "$ARGUMENTS"`
+- иначе → `claudenv decisions search "$ARGUMENTS"`
+
+## Output
+
+Печатай stdout команды без редактирования. Если команда вернула non-zero exit code — покажи stderr пользователю и предложи `claudenv decisions list` для проверки доступных id.
+
+## Если `claudenv` не на PATH
+
+Подскажи: `npm i -g claudenv` или `npx claudenv decisions list`.

package/scaffold/global/.claude/commands/deeper.md

@@ -0,0 +1,55 @@
+---
+description: Развернуть последнее принятое решение в подробное объяснение (deep dive — концепт, варианты, канон)
+allowed-tools: Read, Glob, Grep, WebSearch, WebFetch, Bash(claudenv:*), Bash(ls:*), Write, Edit
+argument-hint: [topic-or-decision-id]
+---
+
+# /deeper — Подробное объяснение последнего решения
+
+Развернуть последнее (или указанное в `$ARGUMENTS`) принятое vibe-decision в полный формат.
+
+## Step 1 — Find the target decision
+
+If `$ARGUMENTS` is empty:
+
+1. Run `claudenv decisions list --limit 1` (Bash) — get the most recent decision id
+2. Run `claudenv decisions show <id>` to load its frontmatter
+
+If `$ARGUMENTS` is a topic slug or id:
+
+1. Run `claudenv decisions show $ARGUMENTS` to load it
+2. If not found — fall back to `claudenv decisions search "$ARGUMENTS"` and pick the top match
+
+If nothing matches:
+
+```
+Не нашёл недавнего решения. Сначала сделай нетривиальный выбор — vibe-decisions залогирует его. Потом вызови /deeper.
+```
+
+Stop here.
+
+## Step 2 — Expand into the deep dive format
+
+Follow `~/.claude/skills/vibe-decisions/deep-dive-template.md` exactly:
+
+1. **Концептуально** (3-5 lines) — what it is, what problem it solves, name 1-2 alternatives
+2. **Как работает** (5-10 lines, optional code sketch) — simplified model + key invariants
+3. **2-3 варианта реализации** — for each: code sketch + when-suits / when-doesn't
+4. **Канон** — 2-4 references, **first** from `~/.claudenv/memories/canon/index.yaml` by topic; if no match, WebSearch and propose `claudenv canon add` afterward
+
+## Step 3 — Update the decision file
+
+Edit the original decision file to set:
+
+```yaml
+deep_dive_done: yes
+sources_consulted: [<urls and canon ids you cited>]
+```
+
+Preserve the `__VIBE_DECISION__` marker on the last line.
+
+## Style
+
+- User's language
+- Concrete to the actual task, not generic textbook
+- If a sketch is shorter than prose — use the sketch

package/scaffold/global/.claude/commands/just-code.md

@@ -0,0 +1,26 @@
+---
+description: Подавить vibe-decisions overview на следующий ответ — просто пиши код без brief overview
+allowed-tools: 
+argument-hint: 
+---
+
+# /just-code — Подавить vibe overview одноразово
+
+Эффект на ОДИН следующий ответ:
+
+- vibe-decisions skill НЕ выдаёт brief overview / auto-log block
+- Просто пишет код для текущей задачи
+- Decision файл НЕ создаётся для этого ответа
+- На ответ после следующего поведение возвращается к дефолту
+
+## В AUTO-LOG mode (loop)
+
+В loop эта команда фактически no-op для большинства случаев — auto-log не паузит, а просто логирует. `/just-code` пропускает логирование на один шаг. Это полезно когда выбор тривиален но vibe-decisions всё равно зажёгся.
+
+## В INTERACTIVE mode
+
+Пропускает паузу на следующем шаге. Используй когда явно знаешь что хочешь, без нужды в overview.
+
+## После использования
+
+Не нужно ничего откатывать — действие истекает само через один turn. Если хочешь надолго отключить vibe-decisions — отредактируй его SKILL.md trigger criteria.