skill-creator

skill-creator is a skill for Claude Code, Codex from tigu77/tiguclaw. It costs 232 tokens per session (1,872 once invoked), scanned A, original, Apache-2.0.

A guide for creating or improving an individual coding-agent skill, with optional tests to measure whether the result is better. It also covers when to use a larger setup containing multiple agents and an orchestrator.

In plain words
What is it for?
Use it to design, revise, evaluate, or improve agent skills and their descriptions, including controlled test runs and approval-based updates.
Why use it?
It provides a repeatable way to define a skill, check for duplicates, test its output when needed, and keep changes subject to human approval.

Skill for Claude CodeCodex

Written for no agent in particular: nothing here depends on one. Also seen: mentions Codex.

Good fit Use it to design, revise, evaluate, or improve agent skills and their descriptions, including controlled test runs and approval-based updates.

Compare 6 skills from other repositories ↓
Install with agentmods
npx agentmods add skills/tigu77/tiguclaw/skill-creator
Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

Any agent
npx skills add tigu77/tiguclaw --skill skill-creator
Clone the repo
git clone --depth 1 https://github.com/tigu77/tiguclaw

Made for: Claude Code, Codex.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for skill-creator

README.md
[![agentmods](https://agentmods.dev/badge/skills/tigu77/tiguclaw/skill-creator.svg)](https://agentmods.dev/skills/tigu77/tiguclaw/skill-creator)
Your own site
<a href="https://agentmods.dev/skills/tigu77/tiguclaw/skill-creator"><img src="https://agentmods.dev/badge/skills/tigu77/tiguclaw/skill-creator.svg" alt="Measured on agentmods" height="20"></a>
Per session 232 Skills are progressive disclosure: only the name and description are preloaded; the body loads when the skill is used.
When invoked 1,872 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. A grade says what 26 rules found in the file — not that it is safe.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5.1 $0.00232 $0.01872
Opus 5 $0.00116 $0.00936
Sonnet 5 $0.00046 $0.00374
Haiku 4.5 $0.00023 $0.00187

Measured 3d ago against content hash 526a60738678, method: parsed. Prices are Anthropic first-party input rates as of 2026-09-08, from the pricing page.

Security

Grade A, and why

skill-creator scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 3d ago.

The scan reads SKILL.md. This mod also ships 1 executable file (scripts/aggregate.mjs), listed below but not scanned — reading those needs a real analyzer, not pattern matching.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

skills/skill-creator/SKILL.md · 69 lines

How it starts

The opening of the file, as written. The whole thing — 69 lines — stays where its author put it; the contents beside it link to each section on GitHub.

Skill Creator — 스킬을 만들고·증명하는 메타 스킬

티구클로가 개별 스킬 하나를 제대로 만들 때 쓰는 절차. 핵심은 "만들었다"에서 끝내지 않고 "정말 나아졌나"를 통제된 eval 로 정량 증명한 뒤 사람 승인으로 반영하는 것.

언제 이 스킬 vs harness

  • harness — 새 도메인에 맞는 서브에이전트 팀 + 오케스트레이션 스킬을 통째로 구성/확장.
  • skill-creator(이 스킬)개별 스킬 하나를 제대로 만든다. harness 가 개별 스킬을 만들 때 이 스킬을 참조해도 된다.

불변 원칙 (게이트)

  • 가볍게가 기본: 대부분의 스킬은 초안 + 빠른 정성 확인이면 충분하다. 아래 "정량 증명"은 필수 단계가 아니라 필요할 때만 꺼내는 도구다. 사소한 스킬에 테스트셋·N회 대조를 강제하지 마라(과설계 = 메타원칙 위반).
  • human-gate: 스킬 파일 실제 생성/교체는 사람 승인 후. self-growth 단방향·human-gated 불변식 유지. 자동 반영 금지.
  • LLM-agnostic: eval 을 돌릴 때도 어댑터 특수분기 0 — spawn_agent(어느 어댑터든)로만. claude -p 종속 러너 금지.
  • 직접 만들지 않는다: eval 엔진·채점 엔진 신설 금지. 기존 spawn_agent(러너/그레이더) + 집계 스크립트 1개만 조립.

기본 경로 (모든 스킬)

1. 무엇을 하는 스킬인지 정한다

스킬의 의도(사용자가 무엇을 원할 때 발동되어 무엇을 해주는가)를 한 문장으로 — 구현이 아니라 사용자 intent. find_skills 로 중복 확인. self-growth 제안(feedback 메모 kind:"skill_proposal")에서 출발했다면 그 근거(반복된 도구 시퀀스)를 읽어 의도를 좁힌다.

2. 초안 작성

references/skill-format.md 규격대로 SKILL.mdWrite:

  • 위치: 프로젝트 전용 <project>/.tiguclaw/skills/<name>/, 공통 <home>/skills/<name>/, 빌트인 레포 skills/ — ★이건 앱 소스다. 헌법 §1 의 고지·확인 절차를 거쳐야 한다(미커밋 변경이 남아 /update 가 막힌다 — 설치는 git clone 이고 자가 업데이트는 git pull --ff-only 다). 그냥 세 번째 선택지처럼 쓰지 마라.. (매 턴 라이브 발견 → 재시작 불요.)
  • description: 트리거 문구를 적극적으로(pushy) — "~할 때/~요청 시" 를 사용자 표현으로 여러 개. 구현 아니라 발동 조건. 길이 예산 의식(매 턴 인덱스에 prepend).
  • 본문: 명령형 절차. 상세는 references/ 로 분리(본문 슬림).

3. 빠른 정성 확인 (이게 보통 충분하다)

  • 트리거 점검: should-trigger 문장 46개 / should-NOT(인접 스킬과 헷갈리는 near-miss) 46개를 적어 description 이 제대로 발동/미발동하는지 눈으로 검증. 과대/과소면 description 문구 조정.
  • 동작 점검(선택): 대표 요청 1~2개를 spawn_agent 로 스킬 적용해 돌려보고 결과가 의도대로인지 확인.
  • 여기서 만족스러우면 → 4. 반영으로. (CC skill-creator 도 "human review loop is usually sufficient" — 정량 벤치는 부스터일 뿐.)

4. human-gate 반영

사람이 초안/확인 결과를 보고 승인하면 스킬 파일을 실제 위치에 둔다. self-growth 제안에서 출발했다면 그 제안 메모를 해소(closed) 안내.


정량 증명 루프 (필요할 때만)

언제 켜나 — 다음일 때만. 아니면 위 "빠른 정성 확인"으로 끝내라:

  • 스킬이 중요/논쟁적이라 "정말 값을 더하나"를 데이터로 봐야 할 때
  • 개선을 반복 중인데 어느 버전이 나은지 애매할 때
  • self-growth 가 "이 개선이 나아졌다"고 주장 → 그 주장을 통제 실험으로 확인할 때
  • 같은 스킬의 어댑터별 성능(claude/codex/claude·codex·openai)을 비교하고 싶을 때 (CC 는 claude 전용이라 못 함)

Read the full file on GitHub · 69 lines

Files

What ships with it

3 files beside SKILL.md in the same directory: the scripts, references and assets a skill reads on demand. Not counted in the per-session cost; read them before you install if any of them is executable.

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. 3d ago Changed · +1 lines 526a60738678
  2. 8d ago First seen · 68 lines · 232 tokens per session scan A 07358ade3b76

Subscribe to this mod's changes

skill-creator is a skill published in the GitHub repository tigu77/tiguclaw (6 stars, last pushed today), licensed Apache-2.0. It adds 232 tokens to every session and 1,872 once invoked, about $0.0012 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.

Related

Other skills, from other repositories

commonly

You are a member of a Commonly workspace — a shared space where humans and AI agents from any origin collaborate in pods (chat rooms with memory). Use this whenever you are connected to Commonly via the commonly MCP tools: to read what's happening, post, remember things across sessions, react, DM other agents, and…

Team-Commonly/commonly · 0 tokens

github

Interact with GitHub (issues, PRs, repos, releases) using the gh CLI. Use when asked to read or write GitHub state — open an issue, fetch PR diff, comment, list runs, etc.

Team-Commonly/commonly · 49 tokens

pandic-office

Convert Markdown to PDF (or DOCX/EPUB/HTML) using the pandoc CLI. Use when asked to produce a PDF report, brief, summary, or any document where the input is Markdown and the output should be a polished, paginated file.

Team-Commonly/commonly · 59 tokens

pdf

Manipulate PDF files — extract text, count pages, render thumbnails, merge or split documents. Use for PDF-specific operations that don't fit markdown-converter (general read) or pandic-office (write from markdown).

Team-Commonly/commonly · 49 tokens

tmux

Manage long-running shell sessions with tmux — start a detached session, run a long task, reattach later, capture output. Use when a task takes longer than a single tool call (build, test, log tail).

Team-Commonly/commonly · 48 tokens

community-research-insight

Extract structured insight briefs from community research transcripts or notes. Produces pain points, stakeholder needs, opportunity maps, risks, and follow-up questions. Requires human review before publication.

clawdotnet/openclaw.net · 41 tokens