菜单
Renders the FH hub map on demand — the door menu, a starter set of skills, and the most-used trigger phrases — without requiring a greeting. State-aware; composes live candidates from the session card and tracks.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
Routes a proposed skill, plugin, or agent to its correct home — forge-harness (FH) meta-skill, project-local agent, or drop — by applying a 4-criteria meta-skill bar followed by a project-local value test.
Reduces token waste in Claude Code sessions across two axes — context footprint (auto-generates .claudeignore, guides over-read files and /clear timing) and command output (routes to a command-output proxy/hook such as rtk to trim verbose CLI stdout). In hub environments, regularly audits bloated CLAUDE.md/MEMORY.md/memory/*.md files and proposes compression. Usable standalone without a hub clone.
Mines improvement insights from long-form AI dialogue logs and long texts — an external or pasted corpus, not the current session (current-session wrap-up belongs to harvest-loop). First operation is sycophancy removal — separates speakers, strips the counterpart's agreement/restatement spans, converts the user's own utterances into falsifiable propositions, and labels each proposition induced (frame first introduced by the counterpart) or independent (originated with the user). Ships a drop counter so removed material is visible, never silently lost. Triggered by "mine insights from this chat log", "이 대화록에서 통찰 캐줘", "strip the agreement and show me what's actually mine", "동조 걷어내고 알맹이만", "what did I actually contribute in this thread".
Collects the latest AI/harness engineering trends from HackerNews and arxiv, then synthesizes them into actionable insights directly relevant to forge-harness operations. Triggered by "AI trends", "latest trends", "this week's AI updates", "what's on the frontier", "harness trends", "frontier-digest".
Scans a project's harness structure (.claude/ directory, CLAUDE.md, rules, agents) to diagnose complexity, drift, missing files, and broken references, then suggests improvements. `--lint` flag adds language-pattern scan (self-marketing, cushion words, version brags). Triggered by "harness check", "structure diagnosis", "CLAUDE.md audit".
Checks a submitted PR against the environment's baseline assets (CLAUDE.md, memory, naming, asset classification) and attaches a review comment with a merge recommendation. 5 steps — diff read, 8-area consistency check, self-catch, comment, merge recommendation.
| name | fh |
| description | Renders the FH hub map on demand — the door menu, a starter set of skills, and the most-used trigger phrases — without requiring a greeting. State-aware; composes live candidates from the session card and tracks. |
| user-invocable | true |
The greeting flow (CLAUDE.md §Active Onboarding) fires on greetings, start intents, new-task and discovery utterances — but it is salience-dependent, once-per-session, and skipped entirely when the user opens with a task. This command is the explicit, deterministic route to the same map: slash autocomplete discoverability, invocable mid-session any number of times, no reliance on the model catching a phrase. Same map, different guarantee — /fh does not claim a gap in which utterances fire onboarding; it closes the how-reliably-and-when gap.
Run the same mechanical branch test as §Active Onboarding: session files / mapped project tracks
under tracks/ (underscore dirs don't count) → new / returning; FH-dev state (session card ·
open fh_signal_* · CLAUDE.local.md) → operator. Do not invent a separate test — the canonical
branch rules live in CLAUDE.md §Active Onboarding and fh_detail_protocols.md Step 2.
Output the door skeleton for the detected branch verbatim from the canonical source (CLAUDE.md §Active Onboarding — including the 🐿️ same-line welcome). Compose door ③ / 🔧 candidates live from the session card and CATALOG, exactly as the greeting path would.
templates/starter_profile.md (read it — do not
hardcode a list that can go stale), one line each.End with "pick a door, say a phrase, or just state your task". Do not auto-run anything — this command is a map, not a dispatcher.
| Condition | Check class |
|---|---|
| Door menu rendered for the correct state branch (new/returning/operator) | mandatory-pass (output exists; branch test is the mechanical §Active Onboarding rule) |
| Menu text matches the canonical skeleton (no drifted fork of the door labels) | measured — at render time, diff the rendered labels against CLAUDE.md §Active Onboarding (the render-vs-source diff IS the check; the canonical-side 4-axis guard only protects the source, not this skill's rendering) |
| Starter set and phrases sourced from their canonical files, not hardcoded | judged — paired with /phantom-quench back-trace (each rendered item must exist in its source file) |
/fh (primary — explicit slash command)Natural-language triggers deliberately overlap the §Active Onboarding discovery triggers — both routes render the same map from the same canonical source, so whichever route catches first, the outcome is identical (collision-safe by construction, not by luck). Baseline Step 0.5 trigger-probe: due at the next harness-doctor run (this skill is a routing surface — obligation per CLAUDE.md §New Skill Creation Pre-Commit Gate).