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 yan-labs/yan-skills --skill skillgit clone --depth 1 https://github.com/yan-labs/yan-skillsWrote 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/yan-labs/yan-skills/skill)<a href="https://agentmods.dev/skills/yan-labs/yan-skills/skill"><img src="https://agentmods.dev/badge/skills/yan-labs/yan-skills/skill/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/yan-labs/yan-skills/skill"><img src="https://agentmods.dev/badge/skills/yan-labs/yan-skills/skill.svg" alt="Reviewed on agentmods" width="80" height="20"></a>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 | $0.00169 | $0.05028 |
| Opus 5.5 | $0.00068 | $0.02011 |
| Sonnet 5.5 | $0.00034 | $0.01006 |
| Haiku 4.5 | $0.00017 | $0.00503 |
Grade A, and why
agent-fleet 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 10d ago.
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 — 159 lines — stays where its author put it; the contents beside it link to each section on GitHub.
agent-fleet
| 让谁做 | 适合什么 | 命令示例 | 派用者顺位参考 |
|---|---|---|---|
gpt |
编码、调研、报告、只读复核；本机 Codex GPT-6.1 Sol | fleet-go new fix --to gpt --auth local --goal "修复问题" --body task.md |
网关 gpt-sol → Grok |
claude |
Claude 月度额度任务；默认 sonnet | fleet-go new check --to claude --tier haiku --goal "复核报告" --body task.md |
code → 网关 gpt-sol → Grok → Gemini（仅文本） |
grok |
视频与图片（生成、搜索）、X 平台（热点与实时讨论）、编码、检索调研、成人题材 | fleet-go new trend --to grok --goal "整理 X 热点" --body task.md |
Claude 顺位下一档 Gemini（仅文本） |
gemini |
文案、翻译、摘要、批量机械任务（--bulk） |
fleet-go new copy --to gemini --goal "写页面文案" --body task.md |
— |
jev |
分类、选择、打分等结构化判断 | fleet-go new decide --to jev --goal "按正文判据分类" --body task.md |
— |
中途插话（所有执行者都能改方向，不要杀掉重派）
统一入口：fleet say <任务> "新指令" 或 fleet-go amend <名字> --say "补充要求"（--restart 只作最后兜底）。
| 执行者 | 插话方式 | 说明 |
|---|---|---|
gpt（Codex，非 --review） |
同一轮里 turn/steer（每个任务自己的 codex app-server） |
失败降级为 SIGINT + codex exec resume（新一轮，同会话），再降级为 --restart；steer 被接受不等于立即停下正在执行的命令，已完成的写入撤不回 |
gpt --review（只读 exec） |
不支持同轮 | 改方向用 amend --restart |
grok |
停掉当前进程，用 grok -r <sessionId> 在同一会话续跑（新一轮，保留上下文） |
拿到 sessionId 之前拒绝，当前任务继续运行 |
claude、gemini（网关与 SDK 任务） |
经收件箱送入运行中的任务 | fleet say 即可 |
jev |
单次请求，无需插话 | |
fleet web（网页版 ChatGPT） |
fleet web say <会话名> "追问" |
|
Agent 工具里的 executor-* |
SendMessage |
不属于 fleet |
fleet team / fleet-go team 一屏检查五产品能力、顺位参考和可用性，只显示登录/凭据存在状态，不显示密钥或 token。
派单前必读 每次用
/agent-fleet派单前、每次 Codex 任务完成后，主线程先跑node ~/Project/kcsx/macmini/yan-skills/agent-fleet/bin/fleet-archive-sweep.mjs，定期扫掉漏归档线程。只归档originator=agent_fleet且已结束、闲置至少 10 分钟的线程；可--dry-run，可用thread/unarchive回滚。运行中的任务仍出现在 Codex APP，结束后归档。 ⓪ 默认用fleet-go new派单（见下「快速派单」）：只写任务独有的正文，归类行、REPORT 行、授权覆盖、已知坑、验收、逐字规则句由它自动补全。不要再手写整份 brief、也不要用 python 拼接旧 brief；手写 brief +fleet code只留给需要完全自定义的少数情形。 ① 唯一正确写法：一条 Bash 调用，命令只含fleet-go new ...、fleet-go relaunch <name>或fleet <子命令> brief.md ...，工具参数run_in_background: true（可重定向到输出文件）。 ② 命令中一律禁止：末尾&、(... &)、> /dev/null 2>&1 &、nohup、setsid、disown、... & sleep、fleet ... && 别的命令 &。 这些写法会让任务脱离工具，没有完成通知，工具任务列表里也看不到。 ③ 补救：发现已经用错，或新会话接手未结束任务，不要杀掉重派；先fleet status --running --json。 对每个 run 各开一条新的 Bash：fleet wait <runId>，工具参数run_in_background: true，重新挂上完成通知。 ④ 启动后自检：工具应返回「Command running in background with ID …」；输出文件持续有内容或任务仍是 running，而不是几秒内就 completed。
What ships with it
27 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.
- evals/evals.json 4.0 KB
- references/codex-coding.md 5.9 KB
- references/copy-voice.md 11 KB
- references/grok-cli.md 5.7 KB
- references/judge.md 851 B
- references/media.md 2.6 KB
- references/operations.md 9.6 KB
- templates/blocks/accept-boilerplate.md 249 B
- templates/blocks/auth-deploy-prod.md 280 B
- templates/blocks/auth-local.md 123 B
- templates/blocks/auth-paid.md 198 B
- templates/blocks/auth-readonly-web.md 151 B
- templates/blocks/auth-review.md 155 B
- templates/blocks/make-image.md 867 B
- templates/blocks/make-video.md 503 B
- templates/blocks/pitfall-fleet-bg.md 122 B
- templates/blocks/pitfall-local-preview.md 162 B
- templates/blocks/pitfall-node-proxy.md 90 B
- templates/blocks/pitfall-playwright-clean.md 161 B
- templates/blocks/pitfall-pnpm-entry.md 105 B
- templates/blocks/pitfall-secrets.md 151 B
- templates/blocks/pitfall-shared-tree.md 197 B
- templates/blocks/pitfall-ssr-check.md 151 B
- templates/blocks/pitfall-zsh.md 87 B
- templates/blocks/report-template.md 741 B
- templates/blocks/rule-sentence.md 327 B
- templates/blocks/style-sol.md 525 B
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.
- 10d ago Changed · +4 lines 1dd47c405054
- 11d ago Changed · -46 lines eb2b4ef24cb2
- 13d ago Changed · +57 lines · +6 tokens per session 6a811d914b5f
- 17d ago Changed · +17 lines 49e6b55d46b6
- 18d ago Changed · +3 lines 3243b9dbbcc7
- 20d ago Changed · +36 lines 4054bc3e482a
- 21d ago Changed 343d28d7511e
- 22d ago Changed · -11 lines · -27 tokens per session 77e72ee5d205
- 23d ago Changed · +28 lines · +28 tokens per session b8c43b19a954
- 24d ago First seen · 71 lines · 162 tokens per session scan A 2a545e53707f
agent-fleet is a skill published in the GitHub repository yan-labs/yan-skills (224 stars, last pushed 5d ago), licensed MIT. It adds 169 tokens to every session and 5,028 once invoked, about $0.0007 per session on Opus 5.5. 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 2026-09-22.
Other skills, from other repositories
seo-mastery-jp
A Japanese-language SEO skill covering search visibility, website structure, page content, structured data, performance metrics, AI search, Astro, and Cloudflare edge deployments.
seo-mastery
Evidence-based SEO and generative search visibility: technical/content SEO, schema.org / JSON-LD, Core Web Vitals, E-E-A-T, Astro, Cloudflare Workers/Pages, and audits. Use for SEO, robots.txt, sitemaps, canonical, hreflang, meta tags, LCP/INP/CLS, GEO (Generative Engine Optimization), LLMO (Large Language Model…
alternatives-pages
Create "[Competitor] alternative" and comparison pages for developer tools. Build honest, high-converting comparison content that ranks for competitive search terms.
analytics-tracking
Design, audit, and improve analytics tracking systems that produce reliable, decision-ready data.
loki-yan-seo
Use when diagnosing or executing Google SEO, international/cross-border sites, AIO, or GEO. Unofficial engine distilled from Loki Yan (@lokiyanseo) public tweets. Not a generic SEO encyclopedia. Does not imitate his voice. Not affiliated or endorsed.
youtube-comment-management
Lister, modérer et répondre aux commentaires de chaînes YouTube via l'API v3, avec maillage SEO vers vos sites WordPress.