agent-init
Skill0 starsChanged 8 months agoMITView on GitHub
Generate AGENTS.md documentation for codebases to guide Code Agent behavior; use when initializing AI coding assistant configurations, standardizing agent behavior, or creating agent guidance docs for new projects
2 other projects use a version of this.
Adopted by
Most recently taken up by
- fritzprix/libr-agenttheir own version, first seen Oct 26, 2026
- dyxushuai/agent-skillstheir own version, first seen Jan 27, 2026
Other versions2 projects use a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 2
- fritzprix/libr-agent★ 11same name
- dyxushuai/agent-skills★ 2same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository3 other files it gives its agents.
- domain-prompt-refinerySkill
- prompt-writerSkill
- prompt-writer-zhSkill
The file
Show its rules and text
We split this skill into the rules it gives an agent
25 rules. About 0 are common, 23 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈23
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 11
Ensure guidelines are stable, high-leverage, and applicable to most coding sessions
- Very specific: line 14
Read the project's key configuration files (package.json, requirements.txt, Cargo.toml, etc.)
- Distinctive: line 40
Read package manager files to identify dependencies and scripts
- Distinctive: line 42
Check for existing documentation (README.md, CONTRIBUTING.md, docs/)
- Distinctive: line 46
Use the template from references/agents-md-template.md to structure the document:
- Distinctive: line 55
Follow the writing guidelines in references/writing-guidelines.md
- Distinctive: line 56
Keep it compact (~60 core lines, not a handbook)
- Distinctive: line 57
Use pointers over copies (progressive disclosure)
- Distinctive: line 58
Avoid linter/formatter details - point to config files instead
- Distinctive: line 59
Every line must justify its permanent presence
- Distinctive: line 70
If essential information is missing:
- Distinctive: line 71
Ask the user 2-4 targeted questions
- Distinctive: line 73
Do not generate incomplete or generic content
- Distinctive: line 84
Avoid edge cases, one-off notes, temporary information
- Distinctive: line 85
Do not turn AGENTS.md into a tutorial or implementation guide
- Distinctive: line 89
Use file:line references for key source files
- Distinctive: line 93
Only include stable, high-leverage information
- Distinctive: line 95
Prefer pointers to detailed documentation over inlining content
- Distinctive: line 99
Do not include style rules (rely on linters/formatters)
- Distinctive: line 103
Before modifying code, read relevant module implementation
- Distinctive: line 105
Follow existing patterns and styles in the codebase
- Distinctive: line 107
When unsure, present options instead of unilateral decisions
- Very specific: line 114
Read package.json to identify dependencies and scripts
- Distinctive: line 116
Check for existing documentation
- Distinctive: line 120
If gaps exist, ask targeted questions
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Files
Agent Initialization - AGENTS.md Generator
Task Objective
- Generate a high-quality
AGENTS.mdfile that provides long-term, general-purpose onboarding for coding agents - Capture the project's WHY / WHAT / HOW framework concisely
- Ensure guidelines are stable, high-leverage, and applicable to most coding sessions
Prerequisites
- Read the project's key configuration files (package.json, requirements.txt, Cargo.toml, etc.)
- Identify the project's primary language and tech stack
- Understand the directory structure and main modules
- Locate existing documentation (README.md, docs/, etc.)
Operation Steps
Step 1: Project Analysis
Analyze the codebase to extract:
WHY (Purpose)
- Core business/technical goals
- Primary users/consumers
- Key constraints (business, technical, regulatory)
WHAT (Architecture)
- Main languages, frameworks, runtimes
- High-level directory structure and module responsibilities
- Key dependencies and their purposes
HOW (Workflow)
- Development environment setup
- Build/test/deploy commands
- Common workflows and validation steps
Analysis Techniques:
- Read package manager files to identify dependencies and scripts
- Examine directory structure to understand module organization
- Check for existing documentation (README.md, CONTRIBUTING.md, docs/)
- Identify configuration files for linters, formatters, CI/CD
Step 2: Generate AGENTS.md
Use the template from references/agents-md-template.md to structure the document:
- Project Purpose: 2-3 sentences capturing core goals and constraints
- Tech Stack and Architecture: Brief list with directory/module responsibilities
- How to Work: Essential commands and validation steps
- Task-Specific Docs: Pointers to detailed documentation
- Working Guidelines: 3-7 high-leverage behavioral rules
Critical Principles:
- Follow the writing guidelines in references/writing-guidelines.md
- Keep it compact (~60 core lines, not a handbook)
- Use pointers over copies (progressive disclosure)
- Avoid linter/formatter details - point to config files instead
- Every line must justify its permanent presence
Step 3: Quality Validation
Validate against the checklist in references/quality-checklist.md:
- Content stability and relevance
- Progressive disclosure implementation
- Actionability and specificity
- Appropriate scope (general vs. specific)
Step 4: Information Gap Handling
If essential information is missing:
- Ask the user 2-4 targeted questions
- Focus on WHY/WANT/HOW gaps
- Do not generate incomplete or generic content
Resource Index
- Template: references/agents-md-template.md - Standard structure for AGENTS.md
- Writing Guidelines: references/writing-guidelines.md - Constraints and best practices
- Quality Checklist: references/quality-checklist.md - Validation criteria
Critical Guidelines
Less is More
- Focus on rules applicable in most sessions
- Avoid edge cases, one-off notes, temporary information
- Do not turn AGENTS.md into a tutorial or implementation guide
Progressive Disclosure
- For details (build steps, test strategy, schema), point to existing docs
- Use file:line references for key source files
- Example: "See
agent_docs/building_the_project.mdfor build details"
Stability Over Completeness
- Only include stable, high-leverage information
- Exclude changing status, TODO lists, historical notes
- Prefer pointers to detailed documentation over inlining content
Agent-Appropriate Scope
- AGENTS.md is for agents; README is for humans
- Do not include style rules (rely on linters/formatters)
- Provide commands to run when lint/format errors appear
High-Leverage Guidelines
- Before modifying code, read relevant module implementation
- For larger changes, propose a plan and reference files to edit
- Follow existing patterns and styles in the codebase
- For breaking changes, prioritize adding/updating tests
- When unsure, present options instead of unilateral decisions
Usage Example
Scenario: User requests AGENTS.md for a Node.js/TypeScript project
Execution Flow:
- Read package.json to identify dependencies and scripts
- Examine directory structure (src/, tests/, docs/, etc.)
- Check for existing documentation
- Identify tech stack (Node version, package manager, frameworks)
- Generate AGENTS.md following template structure
- Validate against quality checklist
- If gaps exist, ask targeted questions
Key Outputs:
- Concise project purpose statement
- Tech stack overview with module responsibilities
- 1-3 essential commands for dev/test/build
- Pointers to detailed documentation
- 3-7 high-leverage behavioral guidelines
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

