agent-workbench
AGENTS.md2 starsChanged 6 months agoMITView on GitHub
Tool-agnostic instructions for AI coding agents (Cursor, Copilot, Codex, Aider, …). AGENTS.md is an emerging cross-tool convention; Claude Code users also have CLAUDE.md, which carries the same rules. Keep the two in sync — or make one a thin pointer to the other so you don't maintain rules twice. This repository is Agent Workbench — tools + methodology for driving an AI coding agent safely on a long-lived codebase. Before you change anything: - Reusable core (scripts/, tools/, .claude/hooks/) is…
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 repository21 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- awb-config-guardSkill
- awb-cookSkill
- awb-dead-code-auditSkill
- awb-debugSkill
- awb-external-refSkill
- awb-handoverSkill
- awb-install-and-verifySkill
- awb-lessons-captureSkill
- awb-optimizeSkill
- awb-output-guardSkill
- awb-plan-then-codeSkill
- awb-researchSkill
- awb-reviewSkill
- awb-session-closeSkill
- awb-stress-testSkill
- awb-tddSkill
- awb-uninstallSkill
- awb-using-skillsSkill
- awb-verify-docsSkill
- prompt-refinerSkill
The file
We split this AGENTS.md into the rules it gives an agent
4 rules. About 1 is common, 3 are distinctive, and 0 are very specific to this project.
≈1
Common · estimate
Many files like this have a rule of this kind.
≈3
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
- could be automatic: a setting could enforce it
- Distinctive: line 13
Read CLAUDE.md — the golden rules and project map live there.
- Distinctive: line 14
Run the checks below to confirm a green baseline.
- Common: line 15
Don't commit secrets, identifiers, or absolute machine paths (the leak scanner enforces this).
- Distinctive: line 32
Best-fit, honest about limits, not gospel: Every tool states what it does *not* do.
≈ “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
Tool-agnostic instructions for AI coding agents (Cursor, Copilot, Codex, Aider, …).
AGENTS.mdis an emerging cross-tool convention; Claude Code users also haveCLAUDE.md, which carries the same rules. Keep the two in sync — or make one a thin pointer to the other so you don't maintain rules twice.
TL;DR for an agent landing here
This repository is Agent Workbench — tools + methodology for driving an AI coding agent safely on a long-lived codebase. Before you change anything:
- Read
CLAUDE.md— the golden rules and project map live there. - Run the checks below to confirm a green baseline.
- Don't commit secrets, identifiers, or absolute machine paths (the leak scanner enforces this).
Checks an agent should run
python -m pytest -q # tests must pass
python tools/leak_scan.py . --entropy --fail-on-find --respect-gitignore # zero leaks (+ high-entropy sweep; skips git-ignored files, as CI does)
python tools/invariants.py . # zero new errorsConventions
- Reusable core (
scripts/,tools/,.claude/hooks/) is stdlib-only. Put dependencies inexamples/and tests. - Every tool needs a runnable example and tests. "Readable but unrunnable" is the failure mode this kit fights.
- Plan → implement → test → review for non-trivial changes.
- Best-fit, honest about limits, not gospel. Every tool states what it does not do. The kit's four tenets are canonical in PHILOSOPHY.md — read it before changing a tool's behavior or the project's voice; it is the kit's reason for existing, not a style note.
- No AI co-author trailers in commits.
Why have both AGENTS.md and CLAUDE.md?
Different tools look for different files. Rather than pick a side, this repo keeps a single
source of rules in CLAUDE.md and uses this file as the tool-agnostic entry point that points
to it. When you adapt this for your own project, decide which file is canonical and make the
other defer to it.
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.

