agentleFS
Sign inStart free
mikeyobrienmikeyobrien/rho/AGENTS.md

rho

AGENTS.md372 starsChanged 8 months agoMITView on GitHub

Rho is an agent configuration + extension bundle for pi: persistent memory via brain.jsonl, heartbeat check-ins, and a knowledge vault. Read brain.jsonl at session start for identity, behavior, and active tasks. Conventions in this repo: - One extension per folder: extensions/<name>/index.ts - Shared code lives in extensions/lib/ - Do not create extensions/lib/index.ts or index.js. Pi will treat extensions/lib/ as an extension and try to load it. - Use extensions/lib/mod.ts as the barrel export instead. Rules: - Do not add browser…

The file

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

9 rules. About 0 are common, 5 are distinctive, and 4 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈5

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 4

    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
  1. Very specific:

    Use extensions/lib/mod.ts as the barrel export instead.

    line 44
  2. Very specific:

    Do not add browser .ts files under web/public/ unless a compile step is introduced.

    line 53
  3. Distinctive:

    Keep browser modules explicit (ES module imports), avoid script-order/global coupling.

    line 54
  4. Very specific:

    Keep web//*.ts and web//*.js at <= 500 lines per file.

    line 55
  5. Very specific:

    Start with .agents/summary/index.md for AI-oriented routing to deeper docs.

    line 60
  6. Distinctive:

    Run tests before declaring code done.

    line 63
  7. Distinctive:

    Verify files were actually written/modified.

    line 64
  8. Distinctive:

    Check command exit codes.

    line 65
  9. Distinctive:

    Prefer explicit over implicit.

    line 76

≈ “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

Rho — Agent Operating Principles

Rho is an agent configuration + extension bundle for pi: persistent memory via brain.jsonl, heartbeat check-ins, and a knowledge vault.

Read brain.jsonl at session start for identity, behavior, and active tasks.

Environment

  • OS: Android / Termux (dev)
  • Arch: aarch64
  • Shell: bash
  • Home: /data/data/com.termux/files/home
  • Brain: ~/.rho/brain/brain.jsonl

The Tenets

  1. Fresh Context Is Reliability — Each check-in starts clean. Read brain.jsonl, re-verify state, plan before acting. Don't assume persistence is perfect.
  2. Backpressure Over Prescription — Don't micromanage how; create gates that reject bad work. Tests, verification, clear pass/fail criteria.
  3. The Plan Is Disposable — Regenerating a plan is cheap. Never fight to save a broken plan.
  4. Disk Is State — Files on disk are ground truth. The brain is continuity between sessions.
  5. Steer With Signals, Not Scripts — When something fails, add a learning, a test, or a behavior entry.
  6. Let Agent Agent — User sits on the loop, not in it. Tune like a guitar, don't conduct.

Work Patterns

Extensions Layout (pi loader rules)

Pi discovers extensions as either single files (extensions/*.ts) or one-level-deep directories with an entrypoint (extensions/*/index.ts). It does not recurse further.

Conventions in this repo:

  • One extension per folder: extensions/<name>/index.ts
  • Shared code lives in extensions/lib/
  • Do not create extensions/lib/index.ts or index.js. Pi will treat extensions/lib/ as an extension and try to load it.
  • Use extensions/lib/mod.ts as the barrel export instead.

Web Code Layout (No-Build Split)

Rho-web intentionally uses a no-build split between Node and browser runtime code:

  • web/*.ts = server/runtime code executed by Node (tsx/strip-types)
  • web/public/js/*.js = browser runtime modules served directly

Rules:

  • Do not add browser .ts files under web/public/ unless a compile step is introduced.
  • Keep browser modules explicit (ES module imports), avoid script-order/global coupling.
  • Keep web/**/*.ts and web/**/*.js at <= 500 lines per file.
  • Respect pre-commit backpressure gates: strict Biome on staged .ts/.js (check --error-on-warnings) + staged web 500-line limit (scripts/pre-commit.sh).

Documentation Entry Point

  • Generated architecture/reference docs live in .agents/summary/.
  • Start with .agents/summary/index.md for AI-oriented routing to deeper docs.

Verification Before Declaration

  • Run tests before declaring code done.
  • Verify files were actually written/modified.
  • Check command exit codes.

Tools Philosophy

  • bash: exploration + system interaction
  • read: inspect before editing
  • edit: surgical changes
  • write: new files or full rewrites only

Prefer explicit over implicit. Show your work.

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