agentleFS
Sign inStart free
btrazzinibtrazzini/hippOS/AGENTS.md

hippOS

AGENTS.md2 starsChanged 14 days agoMITView on GitHub

Reader: agent (Claude Code or Codex). English. This repository is the hippOS system (protocol, skills, validator, templates). Each person's memory lives in a folder in their Google Drive, never here.

  • 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

2 rules. About 0 are common, 1 is distinctive, and 1 is 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.

  • 1

    Very specific

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

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

    Never write anyone's data (names of people and clients, amounts, decisions, contacts, real areas) into a file in this repo.

    line 16
  2. Distinctive:

    Commit messages in English.

    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

hippOS

Reader: agent (Claude Code or Codex). English.

This repository is the hippOS system (protocol, skills, validator, templates). Each person's memory lives in a folder in their Google Drive, never here.

If asked for an assistant, sleep or setup

  • /assistant <area> or $assistant <area>: follow skills/assistant/SKILL.md.
  • /sleep [area] or $sleep [area]: follow skills/sleep/SKILL.md.
  • /configure or $configure: follow skills/configure/SKILL.md (setup, new areas).
  • /protect or $protect: follow skills/protect/SKILL.md (what each area's assistant must never do, and the guard hook).
  • Memory rules: system/protocol.md.
  • Outside agent without the skills (e.g., OpenAI's Dot): EXTERNAL-AGENTS.md too.

If asked to change the system

  • Never write anyone's data (names of people and clients, amounts, decisions, contacts, real areas) into a file in this repo. Examples use generic placeholders (work, hobby, Diego). What belongs to one person (hooks in .claude/settings.json, their own skills) stays only in their private copy.
  • Nothing in the system depends on a fixed name, account, folder or timezone: name, timezone and language come from core/PROFILE.md, the root is the hippOS folder in My Drive (memory.py root locally, a search by name, owner and parent through the connector). In the rules, the person is <name>.
  • Single source: system/protocol.md, tools/memory.py, tools/lookup.py and tools/guard.py. In the skills they are symlinks; don't copy them.
  • Rule change: protocol, skills and validator in the same change, and CLAUDE-PROJECT-INSTRUCTIONS.md when it repeats the rule (the release notes then say to paste it again); python3 -m py_compile tools/memory.py tools/lookup.py tools/guard.py; bump the version with tools/version.sh <new> (protocol, skill titles, VERSION in memory.py).
  • Budget: the protocol and skills, which the assistant and sleep read every session, have a word ceiling (tools/version.sh checks it and refuses to bump the version above it; skills/assistant/opening.md, read only in connector mode, stays out). A new rule only comes in with the case that failed, and for one to come in, another leaves or merges. Test for each line: would removing it make the assistant err? What is mechanical goes into code, not text.
  • Outside inspirations: system/inspirations.md. A curious look; nothing comes in without passing the question "does this help an executive?".
  • Each version: one commit titled X.Y.Z: <what>, a tag vX.Y.Z on it and a GitHub release with short notes. The second number changes when the memory format changes (sleep converts it); the third, for everything else.
  • Commit messages in English; push only on request.

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