Menu
Expand the
Instalar com Codex ou Claude Copie este prompt, cole no Codex, Claude ou outro assistente e deixe que ele revise a página da skill e instale para você.
Baseado na classificação ocupacional SOC
Use when the user describes a multi-step feature or refactor and you need to decide whether to handle it solo, dispatch a parallel `/fan-out`, or run a `/worktree-team` planner→generator→reviewer pipeline. Outputs a routing recommendation with the reasoning behind it.
Use whenever the user asks to apply the same operation to many independent files or components — bulk migration, batch refactor across controllers, parallel exploration of unrelated subsystems. Dispatches one subagent per file with a shared prompt template, then collects the results.
Use to toggle the safe-mode flag at .claude/safe-mode. The flag is written automatically by consecutive-failure-guard once tool failures hit FORGE_SAFE_MODE_THRESHOLD (default 5), and block-destructive.sh denies every Bash/Write/Edit while it exists. /safe-mode off clears the flag, logs the exit, and prompts /postmortem; /safe-mode on enters manually; /safe-mode status reports current state.
Run the 5-stage Token Transformation Pipeline (TRAE §5.2.2) over the current session state — Collection, Ranking, Compression, Budgeting, Assembly. Emits a concrete next-action recommendation (/compact, /lean-md, or /progress-log + fresh session) instead of a generic warning.
Audit or scaffold a CLAUDE.md against the Karpathy 4-section structure (Think Before Coding · Simplicity First · Surgical Changes · Goal-Driven Execution). Companion to /lean-md — structure audit, not size trim.
Print every policy enforcement point declared in plugins/diagnostics/registry/policies.json — id, verdict, plugin, hook event, severity, bypass — grouped by verdict. Single discoverable index of what the harness blocks, anchors, nudges, or logs at runtime.
| name | feature-list |
| description | Expand the |
| when_to_use | Reach for this immediately after a plan is approved (ExitPlanMode) and before the generator dispatches — `/tdd-loop` and `/verify` both read the resulting features.json. Do NOT use for free-form work without a plan; feature-list expects a structured `## Contract` section as its source of truth. |
| disable-model-invocation | true |
| allowed-tools | ["Read","Write","Glob","Grep","Bash"] |
| logical | .claude/features.json contains a testable requirements array derived from the plan's contract |
Convert the latest plan's ## Contract into .claude/features.json. This file becomes the shared source of truth for what "done" means; /tdd-loop consumes it, /verify runs the verify_cmds, and after-subagent.sh flips items to status: done.
Locate the active plan. Run bash plugins/workflow/skills/orchestrate/scripts/find-active-plan.sh (the single source of truth — sort -V numeric-prefix order, skips gate-complete plans). If it prints empty, stop: No plan found. Run Plan mode first.
Extract ## Contract. Read the plan; find the ## Contract section. If missing, stop: Plan has no ## Contract section. /feature-list expects one.
Parse criteria. Each bullet under ## Contract becomes one feature. Trim; skip empty lines and sub-bullets (sub-bullets become part of the parent description).
Derive verify_cmd per item. Inspect the criterion text and the repo stack:
<test-runner> --filter=<hint> (pest/pytest/jest/cargo test/go test).test -f <path> or grep -q '<expected>' <path>.<cmd> | grep -q '<pattern>'.# manual — flag for review.Emit .claude/features.json (overwrite if present; this is a fresh expansion per plan):
[
{
"id": "F1",
"description": "<contract bullet>",
"verify_cmd": "<command or '# manual'>",
"status": "pending",
"plan": ".claude/plans/<file>.md"
}
]
IDs are stable: F1, F2, ... in Contract order.
Report: Expanded N features from <plan>. Run /tdd-loop or /verify to execute verify_cmds.
bash plugins/workflow/skills/orchestrate/scripts/find-active-plan.sh (abort if it prints empty)## Contract section (abort if missing)verify_cmd using the criterion text + repo stack (test runner / test -f / grep -q / # manual).claude/features.json with stable IDs F1, F2, … preserving any prior done entriesExpanded N features from <plan>. Run /tdd-loop or /verify to execute verify_cmds..claude/features.json is read by:
/verify (evaluator plugin) — runs each verify_cmd and writes results.after-subagent.sh (workflow plugin) — flips status when matching work completes.surface-progress.sh (this plugin, SessionStart) — summarizes pending/in_progress/done counts./rest-audit — Efficiency axis (pending/done ratio)./tdd-loop does not read this file; it accepts a feature description as $ARGUMENTS and runs Red/Green/Refactor against the test suite. Use /feature-list first to make the contract machine-readable for /verify, then drive /tdd-loop per item..claude/features.json exists and has done items → do NOT clobber silently. Read it; preserve completed entries when re-expanding; assign new IDs to new items.verify_cmd: "# manual" and report count; the user can refine.# manual.