Solar-Ai / templates
AGENTS.md1 starsChanged 8 months agoMITView on GitHub
- Installs packages
Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.
Of the AGENTS.md files we read, the 32,671 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.
In this repository8 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .claude/CLAUDE.mdCLAUDE.md
- vibe-agentsSkill
- vibe-buildSkill
- vibe-prdSkill
- vibe-researchSkill
- vibe-techdesignSkill
- vibe-workflowSkill
The file
We split this AGENTS.md into the rules it gives an agent
9 rules. About 0 are common, 6 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈6
Distinctive · estimate
This author's own choices. Worth a look.
3
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- could be automatic: a setting could enforce it
- Distinctive: line 18
Do NOT modify these areas without explicit human approval:
- Distinctive: line 25
Architecture rules: Use feature-based folder organization.
- Distinctive: line 26
Testing Expectations: All new utilities must have unit tests.
- Very specific: line 27
Type Safety: Use strict TypeScript. Avoid any types.
- Distinctive: line 31
Plan Before Execution: ALWAYS propose a brief step-by-step plan before changing more than one file.
- Distinctive: line 32
Refactor Over Rewrite: Prefer refactoring existing functions incrementally rather than completely rewriting large blocks of code.
- Very specific: line 33
Context Compaction: Write states to MEMORY.md or a spec.md instead of filling context history during long sessions.
- Very specific: line 34
Iterative Verification: Run tests or linters after each logical change.
- Distinctive: line 35
Team Coordination: If working in Agent Teams, require the Team Lead to approve teammate PRs or plans.
≈ “Common” is an estimate: at least 1 in 5 of the 41,581 AGENTS.md 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
AGENTS.md — Master Plan for [App Name]
Project Overview & Stack
App: [App Name] Overview: [One-paragraph description of the project, its core value proposition, and primary users] Stack: [Primary Tech Stack, e.g., Next.js, React, Node, PostgreSQL] Critical Constraints: [e.g., Mobile-first design required, Multi-tenant architecture, Strict TypeScript adherence]
Setup & Commands
Execute these commands for standard development workflows. Do not invent new package manager commands.
- Setup:
[npm install / pnpm install] - Development:
[npm run dev] - Testing:
[npm test] - Linting & Formatting:
[npm run lint] - Build:
[npm run build]
Protected Areas
Do NOT modify these areas without explicit human approval:
- Infrastructure:
infrastructure/, Dockerfiles, and deployment workflows (.github/workflows/). - Database Migrations: Existing migration files.
- Third-Party Integrations: Payment gateway configurations and Auth setups.
Coding Conventions
- Formatting: Enforce required ESLint/Prettier rules strictly. No warnings allowed in new code.
- Architecture rules: Use feature-based folder organization. Maintain hexagonal boundaries where domain logic does not depend on external frameworks.
- Testing Expectations: All new utilities must have unit tests. Core user flows require integration tests.
- Type Safety: Use strict TypeScript. Avoid
anytypes; define precise interfaces or useunknown.
Agent Behaviors
These rules apply across all AI coding assistants (Cursor, Copilot, Claude, Gemini):
- Plan Before Execution: ALWAYS propose a brief step-by-step plan before changing more than one file.
- Refactor Over Rewrite: Prefer refactoring existing functions incrementally rather than completely rewriting large blocks of code.
- Context Compaction: Write states to
MEMORY.mdor aspec.mdinstead of filling context history during long sessions. - Iterative Verification: Run tests or linters after each logical change. Fix errors before proceeding (See
REVIEW-CHECKLIST.md). - Team Coordination: If working in Agent Teams, require the Team Lead to approve teammate PRs or plans.
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.

