agentleFS
Sign inStart free
AIOSAIAIOSAI/AIPass/AGENTS.md

AIPass

AGENTS.md312 starsChanged 10 days agoMITView on GitHub

Multi-agent framework. Autonomous agents (citizens) live in branches, deploy disposable sub-agents to do work. On any greeting, silently run this sequence — no narration, no announcing steps. Just do it and respond with the status. These steps are sequential and dependent — run each ONCE, wait for the result, then proceed. Never batch a command with its own follow-up read, and never fire duplicate calls. If output looks blank, wait — don't retry. Use drone commands for all operations. Never…

  • 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 repository4 other files it gives its agents.

The file

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

4 rules. About 0 are common, 1 is distinctive, and 3 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.

  • 3

    Very specific

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

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

    Read, one file per call, never concatenated: .trinity/passport.json, .trinity/local.json, .trinity/observations.json.

    line 13
  2. Very specific:

    Never read your README.md at startup — it is the face for strangers.

    line 14
  3. Distinctive:

    Use drone commands for all operations.

    line 19
  4. Very specific:

    Update .trinity/ at natural breakpoints, after milestones, and on /memo (Claude Code, OpenCode) or $memo (Codex).

    line 23

≈ “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

AIPass

Multi-agent framework. Autonomous agents (citizens) live in branches, deploy disposable sub-agents to do work.

User: user

Startup protocol

On any greeting, silently run this sequence — no narration, no announcing steps. Just do it and respond with the status.

These steps are sequential and dependent — run each ONCE, wait for the result, then proceed. Never batch a command with its own follow-up read, and never fire duplicate calls. If output looks blank, wait — don't retry.

  • Read, one file per call, never concatenated: .trinity/passport.json, .trinity/local.json, .trinity/observations.json. A "persisted output" notice means you batched: read the originals one per call, never the persisted file.
  • Never read your README.md at startup — it is the face for strangers. Your inventory is drone @<self> (live) and drone @<self> --help.
  • Refresh: drone @prax dashboard refresh @<self> — where <self> is your branch name (CWD directory name)
  • Dashboard: Read DASHBOARD.local.json — act on what needs attention (new mail → check inbox, active plans → note them). This is your single status glance.

Use drone commands for all operations. Never raw git, gh, file access, or python -m when drone provides it.

Memories

Update .trinity/ at natural breakpoints, after milestones, and on /memo (Claude Code, OpenCode) or $memo (Codex).

Todos[] are a sticky-note pad, not a log: 10 live, each {number, date, task, priority?}, the task one line under 100 chars ("Check on sentry's errors in logs"), no status field — the story lives in plans and sessions. Delete each todo the moment it's done. The oldest roll off by number to .backup/todo/<branch>/backlog.json at rollover (PreCompact) and at @memory's push; drone @memory todo backlog reads them, todo restore <n> brings one back. Rolling never closes anything, so reconcile on load: remove what is finished so it never resurfaces as open. A truncated .trinity read is a signal, not noise — say so first, then measure the file.

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