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 agentmods add instructions/treattrick/skill-sync/agents-mdgit clone --depth 1 https://github.com/TreatTrick/skill-syncWhat 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 | $0.02313 | $0.02313 |
| Opus 5 | $0.01156 | $0.01156 |
| Sonnet 5 | $0.00463 | $0.00463 |
| Haiku 4.5 | $0.00231 | $0.00231 |
Grade A, and why
skill-sync AGENTS.md 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 2d 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 — 163 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
General Development Rules
- Prefer existing repository patterns. Use this document when the codebase has no stricter rule.
- Preserve existing comments. If you need to add comments, write them in Chinese.
- Keep changes focused. Do not perform broad unrelated refactors.
- Use
rg/rg --filesfirst when searching. - Use
apply_patchfor manual file edits.
AI Collaboration Rules
- Prefer implementing the user's current request directly. Ask only when missing information makes a safe assumption risky.
- Do not delete user code, comments, or uncommitted changes. If unrelated changes exist, leave them alone.
- If the user explicitly says not to write tests, do not add tests. Otherwise choose the smallest verification that matches the risk.
- After UI, layout, color, or copy changes, run the matching lint command. Before committing, run
npm run lintandnpm run build. - Do not add empty abstractions, thin components, thin helpers, or functions that only forward runes state reads or setters.
- Report changed files and verification commands. Do not claim unverified success.
Git Commit Rules
- Commit messages must follow Conventional Commits, for example
feat: add user filter,fix: handle empty response, ordocs: update usage notes. - Allowed types are
build,chore,ci,docs,feat,fix,refactor,revert, andtest. - Commit subjects must not use Start Case, PascalCase, or all caps.
- The project uses a
huskycommit-msghook that runscommitlint. - The project uses a
huskypre-commithook that runsnpm run precommitbefore each commit. npm run precommitrunslint-stagedand checks only staged files.lint-stagedformats staged files withprettier --ignore-unknown --write, and runs staged source lint checks for matchingsrcfiles; do not hand-format around the configured Prettier rules.
Anti-Thin-Wrapper Rule
- Do not create wrappers that only forward simple runes state reads or setter calls.
- Add a wrapper only when it contains real semantics such as business rules, defaults, validation, unit conversion, error handling, telemetry, cross-state coordination, or stable reuse from 3+ call sites.
- If a function only does
return state.xxxorstate.setXxx(value), inline the state access at the use site. - When reading state only to apply a fallback, prefer keeping the fallback explicit at the use site, for example
const value = stateValue ?? 0. - Before adding any helper, confirm that it adds meaning rather than just another name.
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.
- 2d ago First seen · 163 lines · 2,313 tokens per session scan A 54e8cd633b78
skill-sync AGENTS.md is an instructions file published in the GitHub repository TreatTrick/skill-sync (4 stars, last pushed 12d ago), licensed MIT. It adds 2,313 tokens to every session, about $0.0116 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.
Other instructions, from other repositories
Khaos-Brain AGENTS.md
Instructions for liuyingxuvka/Khaos-Brain, covering repository expectations, bootstrap, start here, purpose and github publish default.
nextclaw AGENTS.md
Instructions for Peiiii/nextclaw, covering nextclaw ai 常驻内核, 产品愿景, 沟通与推进, 深思与目标模式 and 协作与 git 安全.
moonproxy-desktop AGENTS.md
Instructions for MoonProxyHQ/moonproxy-desktop, covering moonproxy, 功能, 技术栈, 目录结构 and 配置与数据存储.
vscode-gitlens CLAUDE.md
Claude Code instructions for gitkraken/vscode-gitlens, a project described as: Supercharge Git inside VS Code and unlock untapped knowledge within each repository — Visualize code authorship at a glance via Git blame annotations and CodeLens, seamlessly navigate and explore Git repositories, gain valuable insights via…
CodePilot AGENTS.md
AGENTS.md instructions for op7418/CodePilot, covering agents.md, 项目协作定位, 共享规则入口(不在此重复,按需读), 自检命令(codex 跑测试时) and codex review 规则.
MangoDisk AGENTS.md
AGENTS.md instructions for harry0703/MangoDisk, covering mangodisk contribution guidelines, product and architecture, implementation principles, naming and text and repository hygiene.