MANDATORY for every coding agent (Claude Code, Codex, or any other) — keep this repository's documentation in sync after any change to behavior, configuration, interfaces, events, schema, or features. Use automatically at the end of change-sets that alter an env var, event type, hook or state behavior, API, schema, WebSocket message, MCP tool, CLI command, or user-facing feature, and whenever the user asks to update documentation.
インストールする前に、エージェントに与えられる指示の中身を確認できます。
This repository keeps an unusually large, multi-surface, multi-language doc set. Docs drift silently because a change often belongs in 6–10 files across 5 languages plus two HTML pages. This skill encodes which docs exist, which change-types touch which docs, and how to propagate consistently (including the wiki i18n + cache-bump dance).
Authoritative inventory with exact section anchors lives in references/doc-map.md — read it when deciding where a specific change lands. The repo rules .claude/rules/docs-markdown.md ("update all affected docs together"), .claude/rules/wiki-i18n.md, and .claude/rules/i18n-parity.md are binding. Translation propagation — the mirrored READMEs, the wiki bundles, and the UI keys — is owned by the i18n-parity skill; run it whenever this skill's mapping sends you into a localized file.
Update docs in the same change-set (PR/commit) as the code, before claiming done — do not wait for the user to ask — whenever the change is observable from outside the module:
.env.example.events.event_type value) → every event-type list/table.Do NOT auto-update for: pure internal refactors with no observable/interface/config change, test-only changes, comment/typo fixes, or work the user explicitly scoped as "no docs". When unsure whether a change is observable, check the mapping below; if it touches any row, update.
| Change type | Docs to update |
|---|---|
| Env var | README-EN.md, README-CN.md, README-VN.md, README-KO.md, README-ES.md (env tables), ARCHITECTURE.md (inline), server/README.md, wiki/index.html (env table) + wiki i18n, .env.example |
| Event type | README-EN.md+CN+VN+KO+ES (hook-event table), ARCHITECTURE.md (Event types line), docs/PLUGINS.md, wiki/index.html + i18n, docs/DATABASE.md (if it enumerates types) |
| Hook behavior / state transition | docs/HOOKS.md, state-machine mermaid diagrams in README-EN.md+CN+VN+KO+ES + server/README.md + docs/DATABASE.md + wiki/index.html, ARCHITECTURE.md (hooks.js row) |
| API route / response | docs/API.md, server/README.md (routes), ARCHITECTURE.md (routes row), server/openapi*.js (code) |
| DB schema | docs/DATABASE.md, ARCHITECTURE.md (ERD/schema) |
| WebSocket message | client/README.md (Event Types), server/README.md, wiki/index.html |
| MCP tool | mcp/README.md, docs/MCP.md |
| Feature / page / background service | README-EN.md+CN+VN+KO+ES (feature table + data-flow list), concise README.md only when the capability is headline-level, ARCHITECTURE.md (module table), index.html (landing blurb), wiki/index.html + i18n, server/README.md or client/README.md |
| CLI command / script | README-EN.md commands, concise README.md only for primary entry points, CLAUDE.md / AGENTS.md, INSTALL.md / SETUP.md |
| New language | Run the i18n-parity skill and work through its new-language checklist — it covers docs/I18N.md, client/src/i18n/**, the switchers, format.ts, README-<XX>.md, and the full wiki bundle |
README-EN.md and/or ARCHITECTURE.md. Get the wording right there; it anchors every full translation. Keep README.md as a concise English-only landing page.README-CN.md, README-VN.md, README-KO.md, and README-ES.md: mirror the SAME edits at the corresponding sections. Keep identifiers, env-var names, event names, and code in English; translate only prose. Render "Waiting" as 等待中 (zh) / Đang chờ (vi) / 대기 중 (ko) / En espera (es). Match each file's existing terminology — read the neighboring lines first. The i18n-parity skill owns this propagation and its glossary.index.html: one concise marketing sentence in the most relevant existing feature card — light touch, no new sections.wiki/index.html: add the detailed prose/table/diagram at the length and in the position its neighbours already use (see Match the wiki's existing shape below), then follow .claude/rules/wiki-i18n.md — add zh + vi + ko + es entries for every new English string to wiki/i18n-content.js, then bump the cache: increment CACHE_NAME in wiki/sw.js and the i18n-content.js?v= query string in wiki/index.html. Skipping the cache bump means returning visitors never see the update.server/README.md, client/README.md, and the relevant docs/*.md per the mapping.stateDiagram-v2 block that models it (they are duplicated across README/CN/VN/KO/ES, server/README, docs/DATABASE, wiki). Keep transition labels consistent.The wiki is a designed page, not a changelog: several of its blocks live in fixed-size boxes. A block written at 2–3x the length of its neighbours breaks the layout, so measure before you write, and copy the pattern you find.
#feature-carousel .feature-card) share one
fixed-height box. All 39 of them are a single <p>; the group runs
209–628 characters with a ~494 median, so write to roughly 450–550
characters (~65–80 words) — one paragraph, no lists, no sub-headings, no
exhaustive enumeration of every keybinding and edge case. Say what the feature is and
the two or three things that make it distinctive; the depth belongs in the
feature's own section further down the page..screenshot-caption) run ~150–300 characters:
the emoji, the bolded screen name, an em dash, one dense sentence.<section> with h3 + <ul>, never
crammed into a card or caption..claude/skills/update-project-docs/scripts/wiki-block-lengths.sh and
confirm your block sits inside the group's budget before finishing.script.js (see
.claude/rules/wiki-i18n.md) — the scroll-reveal pass adds classes at
runtime, so a :not([class]) selector that looks right in the file and in
jsdom can apply to nothing on the live page.scripts/doc-coverage.sh <new-term> [...] (e.g. the new env var / event type / identifier) and confirm every doc the mapping flags shows a HIT. The matrix is advisory — not every term belongs in every file — but a flagged doc reading 0 is a miss to fix.source --> target: label).zh, vi, ko, and es; cache versions bumped — CACHE_NAME in wiki/sw.js plus the ?v= query strings, which must match between wiki/index.html and the service-worker PRECACHE list (it matches on the full URL, query included, so a stale entry is simply never served)..claude/skills/update-project-docs/scripts/wiki-block-lengths.sh exits 0 — no carousel card or caption is an outlier, and any new card is placed by importance rather than dropped at the front.npm run format (or prettier --check on touched files); for any code touched, run the verification from CLAUDE.md (npm run test:server / test:client / mcp:typecheck).grep -n "<existing-neighbor-term>" <doc> (e.g. grep an adjacent env var to find the env table). references/doc-map.md lists the stable anchors per file.まだレビューはありません。使ってみた感想をお寄せください。
概要と使いどころ
Inspect fired CCAM alerts and manage alert rules for token thresholds, event patterns, inactivity, and status duration. Use when acknowledging alerts, creating or editing a rule, checking cooldowns, or connecting alert rules to webhook targets.
日本語の概要は準備中です。原文の説明を表示しています。
Identify anomalous sessions using Agent Monitor data — cost outliers from the pricing engine, token anomalies (cache miss spikes, compaction baseline surges), unusual event type ratios (PreToolUse/PostToolUse gaps, APIError clusters), behavioral deviations from workflow intelligence (complexity score outliers, error propagation anomalies), and sessions with abnormal metadata (extreme turn_count, high thinking_blocks, zero turn_duration).
日本語の概要は準備中です。原文の説明を表示しています。
Produce a detailed report on APIError events from Agent Monitor data — counts over time, which sessions and models are affected, and the likely root cause (rate limits, overload/529, or context-window pressure) inferred from each event's summary and data payload. Use when API errors spike or when you need to explain why requests are failing.
日本語の概要は準備中です。原文の説明を表示しています。
Benchmark one session (or a small recent set) against the rolling average using Agent Monitor data — cost, total tokens, tool count, and workflow complexity score — and report where each metric lands as a percentile of the population. Tells you whether a session was normal, cheap, or an outlier. Use when judging whether a session was typical or out of band.
日本語の概要は準備中です。原文の説明を表示しています。
Define a spend budget for Claude Code and, optionally, create a cost alert rule that fires when usage crosses the limit, via POST /api/alerts/rules on the Agent Monitor dashboard. Reads current spend from /api/pricing/cost to size the budget sensibly and explains every rule field before writing. Use when setting a spend limit or wiring up a budget guardrail.
日本語の概要は準備中です。原文の説明を表示しています。
Analyze prompt-cache effectiveness for Claude Code usage from the Agent Monitor dashboard — cache hit rate (total_cache_read / (total_cache_read + total_input)), cache_write vs cache_read reuse, cache-read vs cache-write spend, and the sessions with the poorest reuse. Pulls token totals from /api/analytics, per-session detail from /api/sessions, and dollar splits from /api/pricing/cost. Use when diagnosing cache spend or deciding whether prompt caching is paying off.
日本語の概要は準備中です。原文の説明を表示しています。