ModelPeer AGENTS.md

Repository instructions for Model Peer, a Bash tool that asks other coding-agent command-line tools for independent, read-only reviews of code changes. CLAUDE.md and GEMINI.md are linked copies of the same instructions file.

In plain words
What is it for?
Use them when modifying Model Peer, its installation script, tests, documentation site, or commands that consult Claude, Codex, or Gemini.
Why use it?
They explain how to keep the three instruction files synchronized and how to test, lint, and maintain a dependency-free command-line tool.

Instructions file for CodexOpenCode

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add instructions/makedirectory/modelpeer/agents-md
Clone the repo
git clone --depth 1 https://github.com/makedirectory/ModelPeer

Made for: Codex, OpenCode.

Per session 4,508 This file is loaded in full into every session.
When invoked 4,508 The same file — it is already loaded in full.
Security scan A 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.04508 $0.04508
Opus 5 $0.02254 $0.02254
Sonnet 5 $0.00902 $0.00902
Haiku 4.5 $0.00451 $0.00451

Measured yesterday against content hash ecc1ca492d4e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

ModelPeer AGENTS.md scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

`install.sh` must work standalone when piped from `curl`, so it carries a verbatim
AGENTS.md · 351 lines

How it starts

The opening of the file, as written. The whole thing — 351 lines — stays where its author put it; the contents beside it link to each section on GitHub.

AGENTS.md

This file provides guidance to coding agents working in this repository. It is the single source of truth for all three ecosystems: CLAUDE.md (Claude Code) and GEMINI.md (Gemini CLI) are symlinks to this file. Edit AGENTS.md; never replace a symlink with a copy.

What this is

Model Peer is a cross-model peer-review bridge. It lets one coding agent consult a different vendor's CLI (claude, codex, gemini) as an independent, read-only engineering peer, and fans a Git diff out to several models for independent review before a synthesizer reconciles the findings.

There is no build step, no dependencies, and no runtime beyond Bash and the vendor CLIs. The entire product is one Bash script.

Commands

make test     # smoke tests against stub CLIs; runs check-sync first
make lint     # bash -n over every script, plus shellcheck when installed
make sync     # regenerate install.sh's embedded copy of bin/model-peer

The docs site lives in documentation/ (Docusaurus, deployed to GitHub Pages by .github/workflows/docs.yml). It is the only part of the repo with Node dependencies; the tool itself stays dependency-free.

It is served from the custom domain https://modelpeer.app, which depends on three things staying in agreement: documentation/static/CNAME (Docusaurus copies it verbatim into the build, and GitHub Pages reads it), and url / baseUrl in docusaurus.config.js. Reverting baseUrl to /ModelPeer/ breaks every asset path on the live site. Link to docs pages from the README as https://modelpeer.app/<page>, never as a github.io URL.

cd documentation && npm install && npm run build   # build fails on broken links

User-facing behavior changes need the matching page updated under documentation/docs/, not only the README. The README is the entry point; the site is the reference.

tests/smoke.sh is a single linear script with no test-case selection. To run one assertion in isolation, copy its block into a scratch file that reuses the stub setup from the top of tests/smoke.sh. The stubs log invocations to $MODEL_PEER_TEST_LOG and never contact a real model or consume usage.

Read the full file on GitHub · 351 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 351 lines · 4,508 tokens per session scan A ecc1ca492d4e

Subscribe to this mod's changes

ModelPeer AGENTS.md is an instructions file published in the GitHub repository makedirectory/ModelPeer (1 stars, last pushed 15d ago), licensed MIT. It adds 4,508 tokens to every session, about $0.0225 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.