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.
npx skills add cookyman74/my_harness --skill myharnessgit clone --depth 1 https://github.com/cookyman74/my_harnessWrote 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.
[](https://agentmods.dev/skills/cookyman74/my_harness/myharness)<a href="https://agentmods.dev/skills/cookyman74/my_harness/myharness"><img src="https://agentmods.dev/badge/skills/cookyman74/my_harness/myharness/github.svg" alt="Measured on agentmods" height="20"></a>Or the 80×15 button, for a site that already has a row of RSS and ATOM ones. Only the verdict fits; the numbers stay here.
<a href="https://agentmods.dev/skills/cookyman74/my_harness/myharness"><img src="https://agentmods.dev/badge/skills/cookyman74/my_harness/myharness.svg" alt="Reviewed on agentmods" width="80" height="20"></a>- NVIDIA SkillSpector warn
SkillSpector: 2 findings, up to medium
These are SkillSpector’s own severities. On a checked sample its high-severity flags on skills were ~96% false positives — a documented command, a public API, a “never do X” rule — so we show them as a caution to read, not a verdict. Why →
- medium analysis-evasion · line 1 Suspicious Unicode normalization or mixed-script contentFix: Review the flagged content for security risks. Ensure no credentials, secrets, or sensitive data are exposed.
- medium Agent Snooping · line 205 Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.Fix: Remove all code or instructions that list or read other skills' files or directories. Skills should operate independently; cross-skill access is a privilege escalation.
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.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00241 | $0.13275 |
| Opus 5 | $0.00120 | $0.06638 |
| Sonnet 5 | $0.00048 | $0.02655 |
| Haiku 4.5 | $0.00024 | $0.01328 |
Grade A, and why
myharness 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 today.
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.
How it starts
The opening of the file, as written. The whole thing — 501 lines — stays where its author put it; the contents beside it link to each section on GitHub.
Harness — The Team-Architecture Factory
핵심 원칙:
- 에이전트 정의(
.claude/agents/)와 스킬(.claude/skills/)을 생성한다. - 에이전트 팀을 기본 실행 모드로 사용한다.
- CLAUDE.md(+ Codex는 AGENTS.md)에 하네스 포인터를 등록한다. — 새 세션에서 오케스트레이터 스킬이 트리거되도록 최소한의 포인터(트리거 규칙 + 변경 이력)만 기록한다. (듀얼 출력은 원칙 8·Phase 5-4)
- 하네스는 고정물이 아니라 진화하는 시스템이다. — 매 실행 후 피드백을 반영하고, 에이전트·스킬·CLAUDE.md를 지속 갱신한다.
- 품질 게이트 2층 (코드/설계 도메인). 내부 생성-검증(같은 세션 QA)과 외부 리뷰 루프(외부 독립 AI 검증, 러너 엔진 제외 — 같은 엔진=같은 맹점)를 병행한다. 같은 컨텍스트 QA는 같은 맹점을 공유하므로 외부 독립 관점이 추가 결함을 잡는다. 단 합의=정답 아님 — 판정 권위는 오케스트레이터. 상세:
references/external-review-loop.md. - 생성물에 교리 주입. 빌더·수정·QA 에이전트의 작업 원칙에 개발 규칙·TDD 교리를 타겟상대 실경로로 주입한다(
[[ ]]·플러그인 내부 경로 금지 — 서브에이전트가 해소 못 함). 상세:references/dev-rules.md,references/tdd-doctrine.md. - 리스크 등급으로 게이트 강도 조절. 무차별 게이트는 과의식이다. 단계마다 경량/표준/중대 등급을 정해 게이트 강도를 맞춘다 (Phase 5-6).
- 듀얼 런타임 (Claude Code + Codex). 두 런타임 거의 대칭(둘 다 skills·agents·MCP·hooks). SKILL.md 포맷 동일이라 정본 공유, 어댑터로 분기할 것만: 진입점(plugin.json+CLAUDE.md / AGENTS.md), 스킬 경로(
.claude/skills//.agents/skills/), 에이전트(.md/.codex/agents/*.toml), 오케스트레이션(Agent 팀원 spawn / Codex subagents·subprocess). 생성 시 양쪽 출력. 상세·검증:references/runtime-adapters.md.
항법 (최소 경로 — 기본은 슬림)
팩토리가 커졌어도 기본은 슬림. 게이트·평가·자기개선은 리스크가 오를 때만 켠다(단순 하네스 강제 금지).
| 하네스 유형 | 반드시 | 생략 |
|---|---|---|
| 단순/비코드 · 코드 경량 | Phase 0~7 코어 + 내부 QA (dev-rules만 선택) | 외부 리뷰·TDD·평가·self-improve |
| 코드/설계 표준 | + 외부 리뷰 1회 + dev-rules·tdd 주입 | 자동 환류 |
| 코드/설계 중대 | + 단계마다 외부 리뷰 + 승인 사다리 + scorecard | 자동 채택(실험) |
최소 경로·구현 상태(active vs 🧪실험/📐설계만)·루프 개요 지도는
references/factory-map.md. 🧪/📐 기능은 생성 하네스가 자동 실행 안 함.
워크플로우
Phase 0: 현황 감사
하네스 스킬이 트리거되면 가장 먼저 기존 하네스 현황을 확인한다.
프로젝트/.claude/agents/·skills/·CLAUDE.md를 읽는다. 듀얼 런타임이면AGENTS.md·.agents/skills/·.codex/agents/도 읽어 양쪽 drift 점검- 현황에 따라 실행 모드를 분기한다:
- 신규 구축: 에이전트/스킬 디렉토리가 없거나 비어있음 → Phase 1부터 전체 실행
- 기존 확장: 기존 하네스가 있고 새 에이전트/스킬 추가 요청 → 아래 Phase 선택 매트릭스에 따라 필요한 Phase만 실행
- 운영/유지보수: 기존 하네스의 감사·수정·동기화 요청 → Phase 7-5 운영/유지보수 워크플로우로 이동
- 업데이트:
update인자 또는 "팩토리 변경 반영"·"빌드된 하네스 업데이트" → Phase 7-7(사용자 수정 보존 재전파)로 이동
What ships with it
28 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.
- references/agent-design-patterns.md 20 KB
- references/behavior-specs.md 8.1 KB
- references/dev-rules.md 3.2 KB
- references/external-review-loop.md 36 KB
- references/factory-map.md 5.0 KB
- references/harness-scorecard.md 4.9 KB
- references/harness-update.md 5.0 KB
- references/loop-self-eval.md 13 KB
- references/orchestrator-template.md 30 KB
- references/qa-agent-guide.md 11 KB
- references/runtime-adapters.md 7.9 KB
- references/self-improvement-loop.md 12 KB
- references/skill-testing-guide.md 11 KB
- references/skill-writing-guide.md 9.6 KB
- references/tdd-doctrine.md 3.3 KB
- references/team-examples.md 13 KB
- references/templates/working-history-skeleton.md 2.9 KB
- scripts/build-scorecard.sh 7.0 KB runs code
- scripts/check-artifacts.sh 5.0 KB runs code
- scripts/check-behaviors.sh 25 KB runs code
- scripts/check-review-tools.sh 127 B runs code
- scripts/emit-loop-scorecard.sh 2.0 KB runs code
- scripts/grade-trajectory.sh 31 KB runs code
- scripts/harness-update.sh 12 KB runs code
- scripts/run-benchmark.sh 25 KB runs code
- scripts/run-policy-audit.sh 7.3 KB runs code
- scripts/run-review.sh 28 KB runs code
- scripts/x.sh 27 B runs code
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.
- today Changed 04ed9af4103a
- 11d ago First seen · 501 lines · 241 tokens per session scan A f05a03a527da
myharness is a skill published in the GitHub repository cookyman74/my_harness (80 stars, last pushed yesterday), licensed Apache-2.0. It adds 241 tokens to every session and 13,275 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 2026-08-30.
Other skills, from other repositories
systematic-debugging
Use when encountering any bug, test failure, or unexpected behavior, before proposing fixes.
local-ai-agents
Build local-first AI agents that run entirely on a developer workstation with Microsoft Foundry Local and Qwen function-calling models. Covers Small Language Models (SLMs), the OpenAI-compatible local endpoint, sandboxed local tools, local RAG with Chroma, local MCP servers, hybrid cloud/local routing, and the…
next-cache-components-adoption
Turn on Cache Components in a Next.js app and resolve the blocking routes it surfaces. Use when the user wants to enable, adopt, or migrate to Cache Components, flip the cacheComponents flag, work through a flood of blocking-prerender / instant validation errors, run the cache-components-instant-false codemod, or…
insight-error-page
Write or audit an insight-kind error page for the Next.js dev overlay. Use when creating a new errors/ .mdx page, auditing an existing one, or checking that a page matches the framework fix cards. Covers page structure, title alignment, FixCard cards with Copy prompt button, code snippets, terminology verification…
next-cache-components-optimizer
Drive a Next.js route to instant navigation by setting up an agentic loop, under Cache Components / PPR, on initial load (hard navigation) and client-side navigation (soft navigation). Encode the goal as a failing @next/playwright instant() e2e and work it to green, one verified route at a time; the shipped test then…
next-partial-prefetching-adoption
Turn on Partial Prefetching in a Next.js app and work through the insights it surfaces. Use when the user wants to enable or adopt Partial Prefetching, flip the partialPrefetching flag, opt routes in with export const prefetch = 'partial', audit Link prefetch={true} behavior, preserve existing prefetched UI with…