Exécutez n'importe quel Skill dans Manus
en un clic

Exécutez n'importe quel Skill dans Manus en un clic

skill-review

Étoiles7

Forks2

Mis à jour5 juillet 2026 à 11:45

Review and promote skill drafts that skill-curator staged from past sessions. Use when the user runs /skill-review, says "review my skill drafts", "promote a staged skill", or asks what skills the self-improvement loop proposed. Lists drafts under ~/.claude/skill-proposals/, runs the writing-skills quality bar on each, then promotes the approved ones into ~/.claude/skills/ or rejects them.

Installation

Installer avec Codex ou Claude Copiez ce prompt, collez-le dans Codex, Claude ou un autre assistant, puis laissez-le vérifier la page du skill et l'installer pour vous.

Exécuter dans Manus

Source

dwarvesf

dwarvesf/dwarves-kit

Ouvrir le dépôt GitHub Voir les dépôts du créateur

Téléchargement

Exécuter dans Manus

Métiers associésSOC

Basé sur la classification professionnelle SOC

Autres occupations informatiquesProfessions informatiques et mathématiques·SOC 15-1299

SKILL.md

readonly

name	skill-review
description	Review and promote skill drafts that skill-curator staged from past sessions. Use when the user runs /skill-review, says "review my skill drafts", "promote a staged skill", or asks what skills the self-improvement loop proposed. Lists drafts under ~/.claude/skill-proposals/, runs the writing-skills quality bar on each, then promotes the approved ones into ~/.claude/skills/ or rejects them.
disable-model-invocation	false

skill-review (the promote gate)

skill-curator's background reviewer stages skill drafts under ~/.claude/skill-proposals/. It NEVER writes ~/.claude/skills/. This skill is the human gate that promotes a vetted draft into the live library. bin/skill-review (from lib/skill-curator/, put it on PATH or invoke by its full path) is the only writer of ~/.claude/skills/.

Flow

List the staged drafts:

skill-review list      # tab-separated: <slug>\t<description>

If "(no staged drafts)", stop and say so.

For each draft, read ~/.claude/skill-proposals/<slug>/SKILL.md and vet it. Run the superpowers:writing-skills checklist (this skill does NOT reimplement it , delegate):
- Is the name class-level (not a PR number, error string, codename, or today-only artifact)?
- Is the description a real trigger (what class of task + when), so a future agent matches it?
- Is it a genuine reusable pattern, or environment-dependent / a complaint that will rot into a refusal? Drop the latter.
- Secret scan: confirm no token / key / credential is in the body. promote also refuses a draft that still contains one, but check first.
- Would this be better as a PATCH to an existing umbrella, or a references/ add, than a new skill? If so, reject and patch by hand.
Decide per draft (always show the user the draft and your read before acting):
- Approve -> skill-review promote <slug> (moves it into ~/.claude/skills/<slug>/; refuses to overwrite a live skill without --force).
- Reject -> skill-review reject <slug> (moves it to _rejected/, recoverable, never deleted).
Report what was promoted vs rejected, and any draft you left staged for the user to decide.

Rules

Never promote a draft you have not read and vetted against writing-skills.
Never --force overwrite a live skill without explicitly confirming with the user first.
Promotion is the ONLY path a draft enters ~/.claude/skills/. The reviewer cannot (it runs with no write tool). Keep it that way.
An optional auto_promote config knob (default OFF) can auto-pass the lowest-risk class only (a references/ add to an existing umbrella) via skill-review auto; everything else is manual.

Plus depuis ce dépôt

même dépôt

observe

dwarvesf/dwarves-kit

Query and render the kit's control plane from an agent session. Use when asked about the fleet's runs, gate verdicts, conformance, spend/tokens/cache economics, the cognitive-debt score, replaying a recorded run, benchmark results, or building/refreshing the observability dashboard. Trigger phrases include "dashboard", "control plane", "how much did we spend", "token usage", "cache hit rate", "cognitive debt", "debt score", "replay run <rid>", "gate log for <rid>", "conformance", "fleet stats", "observability numbers". All surfaces are CLIs under lib/bench; agents should prefer the JSON verbs over parsing HTML.

2026-07-257

memory-tidy

dwarvesf/dwarves-kit

Use when auditing, consolidating, or cleaning a repo's .claude/memory store, "dọn memory", "memory tidy", the biweekly memory audit, duplicate or stale memory notes, MEMORY.md index drift (notes missing from the index, index entries with no file), or when a scheduled run asks for the memory-tidy audit of a repository. NOT for the built-in machine-local auto-memory under ~/.claude/projects (not git-tracked, no PR gate) and NOT for editing a single note (just edit it).

2026-07-207

stats

dwarvesf/dwarves-kit

Query or render the state of the scattered kit/tide/tg-cleanup/learned ledgers (the dwarves-kit gate/proof/telemetry corpus, tide file-move state, tg-cleanup snapshots, the learning ledger), and detect + propose backlog rows off anomalies in that state. Use when the operator asks to SEE ledger state -- "show me the ledger state", "my debt", "understanding debt", "telemetry", "token cost" / "how much am I spending on tokens", "kit runs" / "kit lane telemetry", "ledger status", "render the ledger", "ledger dashboard", "share this as an artifact" (mid-ledger-conversation) -- or to check/propose off it -- "any ledger anomalies", "is my debt over threshold", "propose a backlog row from the ledger state". Drives the read-only `stats` CLI (dwarves-kit/lib/stats) and renders the result as EITHER an in-terminal reply (bot-reply-formatting , tables + bar-fills) or a shareable web Artifact, both from the same one query; `stats anomalies` is the feedback-loop path, PROPOSE-only (stages a cc-backlog candidate, never files

2026-07-187

get-api-docs

dwarvesf/dwarves-kit

Fetch curated API documentation using Context Hub (chub) before coding against any external API. Use when the task involves calling a third-party API (Stripe, OpenAI, Airwallex, Notion, etc.) to prevent hallucinated parameters and deprecated endpoints.

2026-03-297

name	skill-review
description	Review and promote skill drafts that skill-curator staged from past sessions. Use when the user runs /skill-review, says "review my skill drafts", "promote a staged skill", or asks what skills the self-improvement loop proposed. Lists drafts under ~/.claude/skill-proposals/, runs the writing-skills quality bar on each, then promotes the approved ones into ~/.claude/skills/ or rejects them.
disable-model-invocation	false

skill-review (the promote gate)

Flow

List the staged drafts:

skill-review list      # tab-separated: <slug>\t<description>

If "(no staged drafts)", stop and say so.

For each draft, read ~/.claude/skill-proposals/<slug>/SKILL.md and vet it. Run the superpowers:writing-skills checklist (this skill does NOT reimplement it , delegate):
- Is the name class-level (not a PR number, error string, codename, or today-only artifact)?
- Is the description a real trigger (what class of task + when), so a future agent matches it?
- Is it a genuine reusable pattern, or environment-dependent / a complaint that will rot into a refusal? Drop the latter.
- Secret scan: confirm no token / key / credential is in the body. promote also refuses a draft that still contains one, but check first.
- Would this be better as a PATCH to an existing umbrella, or a references/ add, than a new skill? If so, reject and patch by hand.
Decide per draft (always show the user the draft and your read before acting):
- Approve -> skill-review promote <slug> (moves it into ~/.claude/skills/<slug>/; refuses to overwrite a live skill without --force).
- Reject -> skill-review reject <slug> (moves it to _rejected/, recoverable, never deleted).
Report what was promoted vs rejected, and any draft you left staged for the user to decide.

Rules

Never promote a draft you have not read and vetted against writing-skills.
Never --force overwrite a live skill without explicitly confirming with the user first.
Promotion is the ONLY path a draft enters ~/.claude/skills/. The reviewer cannot (it runs with no write tool). Keep it that way.
An optional auto_promote config knob (default OFF) can auto-pass the lowest-risk class only (a references/ add to an existing umbrella) via skill-review auto; everything else is manual.