agentleFS
Sign inStart free
lmagsinolmagsino/ai-light-factory/AGENTS.md

ai-light-factory

AGENTS.md1 starsChanged 42 days agoMITView on GitHub

AI Light Factory (alf): two skills in skills/ (dev-loop and verify, Agent Skills format), seat contracts in skills/dev-loop/seats/, scripts in bin/, git/gh shims in bin/shims/, a guard in hooks/, and templates in templates/. Decisions, plan and tickets come from Matt Pocock's and Addy Osmani's skill packs; don't re-implement them here. It installs as a Claude Code plugin; the seats can run any coding agent. Any agent that reads AGENTS.md reads this file.

  • 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 31,774 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 repository2 other files it gives its agents.

The file

We split this AGENTS.md into the rules it gives an agent

1 rule. About 0 are common, 0 are distinctive, and 1 is very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈0

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 1

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Very specific:

    Prefer a mechanical check to a sentence: If a rule can live in a script (bin/) or the hook, put it there and have the skill call it.

    line 9

≈ “Common” is an estimate: at least 1 in 5 of the 40,246 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.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
AGENTS.md

Working on AI Light Factory

AI Light Factory (alf): two skills in skills/ (dev-loop and verify, Agent Skills format), seat contracts in skills/dev-loop/seats/, scripts in bin/, git/gh shims in bin/shims/, a guard in hooks/, and templates in templates/. Decisions, plan and tickets come from Matt Pocock's and Addy Osmani's skill packs; don't re-implement them here. It installs as a Claude Code plugin; the seats can run any coding agent. Any agent that reads AGENTS.md reads this file.

Rules for changes

  • Skills follow the Agent Skills spec. name must match the directory name: lowercase letters, digits and hyphens, at most 64 characters. description must be at most 1024 characters, and say what the skill does and when to use it. Keep each SKILL.md under 500 lines, and move detail into references/.
  • Seat contracts stay portable. The files in skills/dev-loop/seats/ are read by path by any agent, so they are plain markdown with no agent-specific syntax. Claude-only frontmatter (disable-model-invocation, argument-hint) is fine in dev-loop, which always runs in Claude Code.
  • Prefer a mechanical check to a sentence. If a rule can live in a script (bin/) or the hook, put it there and have the skill call it.
  • Scripts: bash that runs on macOS's bash 3.2 (no ${x,,}, no mapfile, no associative arrays), with set -euo pipefail, and only jq, git, gh and python3 stdlib. They must pass shellcheck. Every script prints usage from its header comment.
  • Every rule cites its reason. A rule with no why gets deleted by the next person who finds it inconvenient.
  • Every factory change cites the run that motivated it, in the PR body: what happened, which stage should have caught it, and the check you added.

Checks

python3 scripts/validate_skills.py       # frontmatter against the spec
shellcheck -x bin/alf-*.sh bin/alf-{seat,dispatch,wait,board,ready,threads,ledger} bin/shims/* scripts/smoke.sh
bash scripts/smoke.sh                     # seats, dispatch, wait, board, limits, ledger, stats and the guard, against fixtures
claude plugin validate . --strict         # when the claude CLI is available

Download now (ZIP)

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
What did you use it for? e.g. Reviewing pull requests in a Rails app

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.

Check your setup, free