agentleFS
Sign inStart free
sageoxsageox/agent-toolkit/AGENTS.md

agent-toolkit

AGENTS.md13 starsChanged 37 days agoMITView on GitHub

ox is SageOx's team-context CLI. It is optional for outside contributors — if you do not have it, skip the line above and carry on; nothing in this repository needs it to build, test, or run. An opinionated toolkit for hosted AI chat agents that run on several chat surfaces at once from one process and identity. Read README.md for the architecture and SETUP.md to run one. The property every change has to preserve: the brain holds no transport or…

  • Installs packages
  • 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 31,542 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 repository8 other files it gives its agents.

The file

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

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

  • 2

    Very specific

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

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

    Run ox agent prime at session start to load SageOx team context — it's fine to read the user's request first.

    line 2
  2. Distinctive:

    Run the CLI from source with the checked-in wrapper — the packages are not published yet:

    line 39
  3. Very specific:

    Ask its question: would this thing still exist, unchanged, if the Buzz adapter were deleted tomorrow?

    line 61

≈ “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
<!-- ox:prime-check -->

Run ox agent prime at session start to load SageOx team context — it's fine to read the user's request first.

AI Agent Instructions

<!-- ox:prime --> Run SageOx `ox agent prime` on session start, after compaction, or clear. (DO NOT REMOVE)

ox is SageOx's team-context CLI. It is optional for outside contributors — if you do not have it, skip the line above and carry on; nothing in this repository needs it to build, test, or run.

Orientation

An opinionated toolkit for hosted AI chat agents that run on several chat surfaces at once from one process and identity. Read README.md for the architecture and SETUP.md to run one.

The property every change has to preserve: the brain holds no transport or write credential. It can only request actions through one guarded chokepoint, and the gateway decides and executes. If a change lets a credential, a relay handle, or a write path reach the brain's process, it is wrong however well it works.

Build and test

Node 22+ and pnpm 10. The vault tests shell out to real age and age-keygen binaries, so install those too — or run mise install, which the checked-in .mise.toml pins.

pnpm install --frozen-lockfile
pnpm test        # vitest
pnpm typecheck   # tsc --noEmit, root and every package

Both must be green before you open a pull request. Changes under deploy/ also need Helm and Terraform; .github/workflows/ci.yml has the exact commands and pinned tool versions.

Run the CLI from source with the checked-in wrapper — the packages are not published yet:

./bin/sageox-agent --help

Layout

PathWhat
packages/core/Gateway, guard, brain seam, MCP broker, job host, memory, policy
packages/adapter-buzz/Buzz (Nostr) surface: relay client, NIP-42, engrams, identity
packages/adapter-slack/Slack surface over Socket Mode
packages/adapter-console/Local console surface
packages/cli/sageox-agent — create, run, doctor, identity, memory, mcp, jobs
deploy/Dockerfile, Compose, Helm chart, Terraform for AWS/EKS
docs/Design specs, RFCs, the setup guide chapters, operational contracts

Conventions

  • Read docs/naming.md before naming anything that leaves this repository — an MCP server, a package, a chart, a service, an environment variable. Ask its question: would this thing still exist, unchanged, if the Buzz adapter were deleted tomorrow? If yes, it is not Buzz's and must not be named after it. test/naming.test.ts fails on the spelling that keeps recurring.
  • Config is parsed as data, never evaluated. The manifest loader must not gain a path that runs what it reads.
  • One change per pull request, with a test that fails without it. See CONTRIBUTING.md.
  • CHANGELOG.md gets an entry under ## [Unreleased] for anything user-facing.

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