groundwork
AGENTS.md0 starsChanged 7 days agoMITView on GitHub
The project conventions live in .github/copilot-instructions.md. Read that file first, all of it. It is the source of truth for every agent working in this repo, whatever tool you are running under. The definition of done and the rules for how to build are there, and are deliberately not copied here, because two copies drift. Procedures that are too long to keep loaded are in .github/skills/<name>/SKILL.md: These are vendored from other repositories, pinned in .github/skills/vendored.json. Each carries a PROVENANCE.md and…
We split this AGENTS.md into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
Distinctive · estimate
This author's own choices. Worth a look.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 4
Read that file first, all of it.
≈ “Common” is an estimate: at least 1 in 5 of the 36,368 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
Agent instructions
The project conventions live in .github/copilot-instructions.md. Read that file first, all of it. It is the source of truth for every agent working in this repo, whatever tool you are running under. The definition of done and the rules for how to build are there, and are deliberately not copied here, because two copies drift.
Procedures that are too long to keep loaded are in .github/skills/<name>/SKILL.md:
| Skill | Load it when |
project-bootstrap | Starting a new project or adding a new component. |
readme-writer | Creating a README, or changing how the project is set up or run. |
spec-first | The request is vague, or the work is bigger than one sitting. |
iac-templating | Writing or reviewing Bicep or Terraform under infra/. |
These are vendored from other repositories, pinned in .github/skills/vendored.json. Each
carries a PROVENANCE.md and the upstream licence. Do not edit them by hand.
| Skill | Load it when | Upstream |
design-taste-frontend | Building a landing page, portfolio or redesign. | Leonxlnx/taste-skill |
swiftui-expert-skill | Writing or reviewing SwiftUI. | AvdLee/SwiftUI-Agent-Skill |
Copilot exposes two more kinds of file through its UI. Other tools can read them as plain procedures:
| Path | What it is |
.github/prompts/*.prompt.md | Step lists behind /new-project, /update-readme, /write-spec and /new-infra. When asked for one of those, follow the file. |
.github/agents/shipwright.agent.md | A build-and-verify persona: build the smallest thing, run it, document from the transcript. |
.github/instructions/*.instructions.md | Rules for a file type, scoped by the applyTo glob in each header. Read the matching one before editing that kind of file. |
docs/review-checklist.md | Walk the diff through it before reporting done. |
Before reporting done:
python3 scripts/readme_check.py
python3 scripts/readme_check.py --text-only
python3 scripts/vendor_skills.py --checkWhere 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 28,693 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.
- CLAUDE.mdCLAUDE.md
- .github/copilot-instructions.mdCopilot instructions
- design-taste-frontendSkill
- iac-templatingSkill
- project-bootstrapSkill
- readme-writerSkill
- spec-firstSkill
- swiftui-expert-skillSkill
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.