PeerHarness / site
llms.txt33 starsChanged 4 days agoApache-2.0View on GitHub
Your team votes on how its AI coding agents behave. You correct your coding agent ("no, use pnpm, not npm"), and PeerHarness turns the correction into a one-sentence rule. Keep it for your own agent, or propose it to your team: once a majority approves, every member's agent follows it. There is no server. What it is, in facts: To set it up, paste the prompt from the README into the coding agent (Claude Code or any of the others…
In this repositoryOne other file it gives its agents.
- AGENTS.mdAGENTS.md
The file
We split this llms.txt into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 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.
0
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Distinctive: line 15
No server and nothing to sign up for.
≈ “Common” is an estimate: at least 1 in 5 of the 3,727 llms.txt 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.
Files
# PeerHarness
> Your team votes on how its AI coding agents behave. You correct your coding agent ("no, use pnpm, not npm"), and PeerHarness turns the correction into a one-sentence rule. Keep it for your own agent, or propose it to your team: once a majority approves, every member's agent follows it. There is no server.
What it is, in facts:
- An open-source command-line tool, written in Rust, under MIT or Apache-2.0 at your option. Source: https://github.com/aryehlev/PeerHarness
- For Claude Code on macOS, and it connects Codex, opencode, Pi, Google Antigravity, DeepSeek Harness, Gemini CLI, Qwen Code, GitHub Copilot CLI, Cursor, Amp, Factory's Droid, Crush, Kilo and goose too. On Linux it works from a terminal, without the confirmation windows. Windows is not supported.
- Early and not audited. Its limits are listed in full, and it has had no security audit.
- It is used by talking to the agent: "From now on, use pnpm here, not npm.", "Keep it.", "Create a PeerHarness team called acme.", "Propose the pnpm rule to acme.", "Vote approve on it."
- Agents suggest, people decide. The agent can note a rule and open a window on the person's screen that shows the exact text. It cannot answer the window: keeping a rule, sharing one, voting, inviting and joining take the person's click. That is not a sandbox.
- A rule is one instruction-only SKILL.md file, an Agent Skill. PeerHarness never executes a rule.
- The vote: one person, one vote, however many devices they have. A rule passes with a majority of everyone entitled to vote. Nobody has a veto, a rejection is recorded with its reason, and an approval cannot be withdrawn. Admitting or removing a member is voted too.
- A rule that passed is installed for every member's agent. Any member can opt out of a single rule on their own machine.
- No server and nothing to sign up for. Each person runs their own node with their own keys, and every node checks every signature and counts every vote for itself.
- Compared with a CLAUDE.md file in git: if the rule is about one repository and the team already reviews pull requests, commit the file. PeerHarness is for rules that follow people across repositories, for groups with no shared repository, and for members who want to see the exact text they vote on and to opt out of a rule alone.
To set it up, paste the prompt from the README into the coding agent (Claude Code or any of the others it connects). The agent installs the release binary, and the person runs `~/.local/bin/peerharness setup` once in their own terminal.
## Documents
- [README](https://raw.githubusercontent.com/aryehlev/PeerHarness/main/README.md): what it does, how to set it up and how to use it in a coding agent
- [PeerHarness from a terminal](https://raw.githubusercontent.com/aryehlev/PeerHarness/main/docs/terminal.md): every step as a command, and how to take it out again
- [Limits](https://raw.githubusercontent.com/aryehlev/PeerHarness/main/docs/limits.md): what it does not do, or not yet
- [Security model](https://raw.githubusercontent.com/aryehlev/PeerHarness/main/docs/security-model.md): threats, guarantees and limits
- [How each part was checked](https://raw.githubusercontent.com/aryehlev/PeerHarness/main/docs/checked.md)
- [Changelog](https://raw.githubusercontent.com/aryehlev/PeerHarness/main/CHANGELOG.md): what each version added
## Optional
- [Architecture](https://raw.githubusercontent.com/aryehlev/PeerHarness/main/docs/architecture.md): the internal contract every crate follows
- [Protocol](https://raw.githubusercontent.com/aryehlev/PeerHarness/main/docs/protocol-v0.md): the wire format, for a second implementation
- [Decisions](https://raw.githubusercontent.com/aryehlev/PeerHarness/main/docs/decisions.md): decisions and why they were made
- [The page for people](https://aryehlev.github.io/PeerHarness/)
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
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.

