Borrowing it
Nothing to install: this file belongs to Shiyao-Huang/awesome-agent-evolution. Take a copy, put it at the same path in your own repository, and replace the rules that are about this project with yours.
curl -O https://raw.githubusercontent.com/Shiyao-Huang/awesome-agent-evolution/main/AGENTS.mdgit clone --depth 1 https://github.com/Shiyao-Huang/awesome-agent-evolutionWrote 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/instructions/shiyao-huang/awesome-agent-evolution/agents-md)<a href="https://agentmods.dev/instructions/shiyao-huang/awesome-agent-evolution/agents-md"><img src="https://agentmods.dev/badge/instructions/shiyao-huang/awesome-agent-evolution/agents-md.svg" alt="Measured on agentmods" 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 | $0.06446 | $0.06446 |
| Opus 5 | $0.03223 | $0.03223 |
| Sonnet 5 | $0.01289 | $0.01289 |
| Haiku 4.5 | $0.00645 | $0.00645 |
Grade C, and why
awesome-agent-evolution AGENTS.md scanned grade C with 1 finding 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 8d 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.
Recursive force deletehighDestructive command
rm -rf with a variable or a broad path is one typo away from removing the wrong tree.
- **禁止执行破坏性 git 操作**：`git reset --mixed HEAD`、`git reset --hard`、`git checkout .`、`git restore .`、`git clean -f`、`rm -rf` 等一律禁止。用户明确指令 2026-05-26。 How it starts
The opening of the file, as written. The whole thing — 311 lines — stays where its author put it; the contents beside it link to each section on GitHub.
AGENTS.md
L1
你在这个仓库的任务是维护 Self Evolve 的研究管线：raw 归 raw，加工归 processed，工作产物归 work，结果归 results，并让全部内容可索引、可追溯、可构建。
L2
先读 docs/project-management/user-direct-inputs.md，再读 CURRENT_GOAL.md：它们只代表用户直接输入，不代表 team 的二手描述。然后读 README.md、CONTENT_INDEX.md、docs/project-management/project-structure.md 和 docs/indexes/master-index.md。任何新增长期产物都要更新索引，任何会影响论文或网站的改动都要跑对应验证。
L3
这个项目的主要矛盾是：素材、分析、论文、网站和结果曾经混在一起，现在必须被同一条数据流管理。不要为了目录看起来干净而破坏已有脚本、论文引用或网站路由；物理迁移前先用 rg 找引用。根目录只留入口、法律文件和兼容文件，新的规则和长文档进入 docs/。项目深度分析要像 model card 一样能教学：说明问题、方法、证据、限制、适用场景和与自进化的关系。完成工作前刷新索引并汇报验证结果。
Current Scale (2026-05-25)
| Layer | Count | Notes |
|---|---|---|
| raw-github | 490 | 100% frontmatter, 22.2% timestamps unknown |
| raw-papers | 198 (128 unique) | 70 duplicate pairs (dash + dot format) |
| raw-social | 650 md + 645 json | 612 numbered pairs + batch files |
| raw-blogs | 655 md + 653 json | 650 numbered pairs |
| raw-social-rank | 234 md | Strict subset of raw-social (seed dataset) |
| projects | 204 model cards | 146 trace to raw-github (29.8% coverage) |
| analysis | 16 | Cross-analysis layer |
| research | 38 | Author profiles, verification audits |
| paper-reviews | 137 | 90.4% match to raw-papers |
| paper-drafts | 12 tex + PDF | 8-chapter survey, 62 bib entries |
| survey | 11 | Chinese-language parallel survey with figures |
| site | 207 pages | Astro static site, built daily |
| reports | 1 | Critical gap — needs consolidation |
Writing Rules
- L0 Highest guidance — public copy quality first: 用户已明确指出，公开文案质量是最高指导。网站、README、README-EN、论文页、topic/blog/SEO 入口和 metadata 不能只是“造了一堆网页”；它们必须有清晰逻辑、可读、可吸收、接近高质量知识型/Hacker News 级别文案，并且中英双语面向全球读者表达同一条证据链。公开页面数量、SEO 覆盖、视觉效果、内部 workflow 都低于这个指导。
- Visualization first: 当图表比文字更容易表达时，必须使用图表（Mermaid DAG、SVG、数据流图）。不要只写文字。
- Mom Test usability: README 和面向用户的内容必须通过 "Mom Test"——一个非专业人士能否理解这个项目做什么、为什么重要？
- Public copy review gate: 所有公开文案，尤其是网页、README、README-EN、SEO/topic/blog 页面和站点 metadata，发布前必须经过
3-5个读者/编辑 agents 与3个学术 agents 的审查。读者/编辑审查关注 Mom Test、可读性、行动路径、中文默认入口和英文镜像；学术审查关注术语准确性、证据链、claim 强度、限制和[UNVERIFIED]标注。无法实际调度 agents 时，必须在交付说明中明确这是缺口，不能把该文案标记为质量已完成。 - Evidence chain: 每个分析结论必须可追溯到 raw 数据源。如果无法追溯，标注
[UNVERIFIED]。 - Layered output: 判断用 1 句话，证据用 3 句话，完整论证用 5+ 句话。
- Audience boundary: README、网站首页、论文页和 SEO 内容是给外部读者/消费者看的，不是 agent 操作手册。不要把启动检查、内部构建命令、agent wiki 流程、handoff 规则或自我系统说明写进 README 主体；这些规则进入
AGENTS.md、CLAUDE.md、CLOUD.md或docs/ops/。 - Consumer-facing README: README 只回答读者关心的问题：这个项目是什么、为什么重要、核心结论是什么、证据在哪里、下一步读什么。内部维护规则可以链接到 agent/cloud 手册，但不要占据 README 叙事。
- Workflow split: 具体分界见 docs/ops/audience-boundary-workflow.md。公开页出现
Workflow A/B、CLAUDE.md Iron Rules、handoff、启动检查、内部构建命令等词时，优先改写成读者能理解的证据边界，并把操作细节移到 ops。 - Bilingual public surface: 网站、README、论文页、SEO/topic/blog 入口和发布 metadata 必须按中英双语治理；中文
zh-CN是默认读者入口，英文是同证据链镜像，不允许用英文替换、稀释或删除中文叙事。
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.
- 8d ago First seen · 311 lines · 6,446 tokens per session scan C 9b54937bd2d0
awesome-agent-evolution AGENTS.md is an instructions file published in the GitHub repository Shiyao-Huang/awesome-agent-evolution (182 stars, last pushed 1mo ago), licensed MIT. It adds 6,446 tokens to every session, about $0.0322 per session on Opus 5. A static security scan graded it C with 1 finding (recursive force delete). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-30.
Other instructions, from other repositories
next.js AGENTS.md
AGENTS.md instructions for vercel/next.js, covering next.js development guide, codebase structure, monorepo overview, core package: packages/next and other important packages.
codex AGENTS.md
AGENTS.md instructions for openai/codex, covering rust/codex-rs, the codex-core crate, code review rules, crate api surface and model visible context.
vscode buildNext.instructions.md
Working notes and architecture documentation for the new esbuild-based build system in build/next. Use when making changes to the new build pipeline (transpile/bundle commands, NLS plugin, source-map handling, resource copying, or self-hosting watch tasks).
vscode oss-third-party-notices.instructions.md
Instructions for microsoft/vscode, covering vs code oss third-party-notices pipeline, architecture, pipeline flow in ci, applying the notice (cutover) and fallback chain (never fail the build).
langchain AGENTS.md
AGENTS.md instructions for langchain-ai/langchain, covering global development guidelines for the langchain monorepo, corridor security analysis, project architecture and context, monorepo structure and development tools & commands.
spec-kit AGENTS.md
AGENTS.md instructions for github/spec-kit, covering agents.md, about spec kit and specify, quickstart — add a new integration in 5 steps, integration architecture and integrationmanifest — file tracking.