agentleFS
Sign inStart free
BeatMagicBeatMagic/acestudio_agent_plugin/AGENTS.md

acestudio_agent_plugin

AGENTS.md4 starsChanged 58 days agoMITView on GitHub

Know-how for agents that operate ACE Studio: the how, not the tools. The tools are the product's CLI and MCP server; this repo carries the skills that tell an agent what to reach for, in what order, and when to hand the work to something else. Install the plugin on Claude Code or Codex; any other harness takes the skills CLI. Install and update mechanics live in README.md. Connecting to ACE Studio is a separate step from installing the skills:…

  • 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 repository5 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:

    Commit messages prefer Conventional Commits (feat:, fix:, docs:, chore:, ...).

    line 25

≈ “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

ACE Studio agent plugin

What this repo is

Know-how for agents that operate ACE Studio: the how, not the tools. The tools are the product's CLI and MCP server; this repo carries the skills that tell an agent what to reach for, in what order, and when to hand the work to something else.

Installing this for a user

Install the plugin on Claude Code or Codex; any other harness takes the skills CLI. Install and update mechanics live in README.md.

Connecting to ACE Studio is a separate step from installing the skills: if you do not see the CLI or MCP tools available to you, point the user at the app's Connect to Agents entry or the External Agent Access docs.

Working on the content

  • Skills are the orchestrator layer: workflow order, when to reach for what, why, decision heuristics, handoff to external tools. CLI/MCP mechanics (syntax, arguments, enums) live in help and get_docs, and a skill points at them rather than restating them.
  • Examples live in the CLI/MCP docs too; a skill carries one only when it is compact and stable.
  • Skill text is harness-agnostic: the SKILL.md convention only, with no Claude Code or Codex specific mechanism.
  • ace-studio-audio-plugins owns plugin control; per-plugin recipes live in its references/ folder, one file per plugin, and land only once verified against the plugin's real UI.

Release and delivery

  • main is the delivery channel: merging ships to installed users, with no version bump and no tags.
  • Ahead-of-product prose (guidance for a Studio build not yet shipped) waits on a branch or PR until the build ships; a skill that needs the newer build says so inline (requires Studio ≥ x).
  • The repo root is the plugin: a root-level commands/, agents/, hooks/ or .mcp.json silently becomes a plugin component.
  • Commit messages prefer Conventional Commits (feat:, fix:, docs:, chore:, ...).
  • Issues here are an inbound funnel for external users; the team ledger is the ACE Studio tracker (BeatMagic/ACE-Studio).

Before editing

Load the writing-for-agents skill before changing any skill or doc here, where the harness has it.

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