repository-agent-setup
Skill177 starsChanged 7 days agoApache-2.0View on GitHub
Add or evolve AGENTS.md and .agents skills, knowledge, and temporary notes while preserving repository rules and GitHub governance.
In this repository19 other files it gives its agents.
- AGENTS.mdAGENTS.md
- .github/copilot-instructions.mdCopilot instructions
- agent-evaluationSkill
- algorithm-developmentSkill
- buildSkill
- change-reviewSkill
- container-buildSkill
- engineering-solution-designSkill
- interface-migrationSkill
- issue-localizationSkill
- lintSkill
- long-horizon-developmentSkill
- model-artifact-integrationSkill
- multi-repo-devSkill
- performance-optimizationSkill
- simulation-regressionSkill
- test-impact-analysisSkill
- unified-core-runtimeSkill
- vendorSkill
The file
We split this skill into the rules it gives an agent
10 rules. About 0 are common, 7 are distinctive, and 3 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈7
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
- Distinctive: line 20
Use when adding agent conventions to a repository or normalizing an existing agent layout.
- Very specific: line 25
Read existing AGENTS.md, provider instructions, .agents/, relevant source and CI commands, and Git status.
- Distinctive: line 31
Keep root AGENTS.md short: working rules, verified command entrypoints, and links to the skill and knowledge indices.
- Very specific: line 33
Keep .github/ for workflows, contribution policies, issue/PR templates, and provider integration.
- Very specific: line 36
Add .agents/skills/<task>/SKILL.md with name and a searchable description, then scope, prerequisites, steps, acceptance, and failure …
- Distinctive: line 40
Keep facts source-backed, narrowly indexed, and separate from procedures.
- Distinctive: line 42
Keep temporary investigations in the session workspace.
- Distinctive: line 46
Do not fabricate architecture, commands, empty capability claims, or duplicate every skill across repositories.
- Distinctive: line 55
Run the configuration check, review entrypoint/command consistency, and compare agent behavior on relevant evaluation cases.
- Distinctive: line 62
Stop for conflicting knowledge ownership, missing command evidence, or local changes that cannot be merged safely.
≈ “Common” is an estimate: at least 1 in 5 of the 372,429 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
Repository Agent Setup
Repository/
AGENTS.md # Shared entrypoint and working rules
.github/ # GitHub governance and provider integration
.agents/
skills/ # How to perform a task
ad_agents/context/ # Existing durable knowledge authority
Session workspace/ # Temporary task state and evidenceWhen
Use when adding agent conventions to a repository or normalizing an existing agent layout. Apply independently to each repository in a multi-repo workspace.
Prerequisites
Read existing AGENTS.md, provider instructions, .agents/, relevant source
and CI commands, and Git status. Identify the durable knowledge authority before
editing.
Steps
- Keep root
AGENTS.mdshort: working rules, verified command entrypoints, and links to the skill and knowledge indices. Do not copy architecture here. - Keep
.github/for workflows, contribution policies, issue/PR templates, and provider integration. Point provider instructions to the shared agent entrypoint; preserve provider-specific rules. - Add
.agents/skills/<task>/SKILL.mdwithnameand a searchabledescription, then scope, prerequisites, steps, acceptance, and failure handling. Include only source-verified commands; keep core steps self-contained. - Link the existing durable authority instead of creating
.agents/knowledge/. Keep facts source-backed, narrowly indexed, and separate from procedures. Preserve existing ownership; do not move established knowledge silently. - Keep temporary investigations in the session workspace. Never store credentials, personal data, or raw sensitive logs. Promote reusable conclusions only when authorized through the existing context-maintenance process.
- Merge existing content; add only useful task skills and truthful indices. Do not fabricate architecture, commands, empty capability claims, or duplicate every skill across repositories.
Acceptance
- Entry links resolve and every advertised skill exists.
- GitHub governance remains intact; shared rules have one authoritative home.
- Existing knowledge remains reachable with explicit migration boundaries.
- Temporary notes are ignored; reusable guidance is tracked in its owning repo.
- Run the configuration check, review entrypoint/command consistency, and compare agent behavior on relevant evaluation cases. Syntax checks do not prove instruction effectiveness.
- Report which repositories changed. Documentation-only setup needs no builds.
Failure handling
Stop for conflicting knowledge ownership, missing command evidence, or local changes that cannot be merged safely. Ask before moving established knowledge, changing executable configuration, or committing/pushing repository changes.
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.

