Skip to main content
agentsbridge

One AGENTS.md. Every AI coding tool. Zero drift.

CI PyPI Python License: MIT Ruff

Bridge · Quick start · Supported tools · CI drift check · FAQ


Why

Your team uses Claude Code. Your teammate uses Cursor. Someone just switched to Codex CLI. And suddenly your coding standards live in five different files that silently drift apart:

AGENTS.md                          ← Codex, Gemini CLI, Jules, Amp, Zed read this
CLAUDE.md                          ← Claude Code reads this (not AGENTS.md)
.github/copilot-instructions.md    ← GitHub Copilot reads this (not AGENTS.md)
.cursor/rules/*.mdc                ← Cursor reads this (not AGENTS.md, older versions)
.windsurfrules / .clinerules       ← Windsurf / Cline read these
CONVENTIONS.md                     ← Aider reads this

agentsbridge fixes this with one command. You maintain one AGENTS.md (the emerging industry standard) — it generates and keeps every tool's file in sync. Edit once, sync everywhere, and let CI catch anything that drifts.

Quick start

pipx install agentsbridge        # or: pip install agentsbridge / uv tool install agentsbridge

Already have rules somewhere? Import them:

agentsbridge import              # seeds AGENTS.md from CLAUDE.md, .cursorrules, etc.

Starting fresh?

agentsbridge init                # creates a starter AGENTS.md

Then — the only command you'll ever need again:

agentsbridge sync
  + created   CLAUDE.md
  + created   .github/copilot-instructions.md
  + created   .cursor/rules/agentsbridge.mdc
  + created   .windsurfrules
  + created   .clinerules
  + created   CONVENTIONS.md
  + created   .roo/rules/agentsbridge.md
  + created   .kilocode/rules/agentsbridge.md
  + created   .junie/guidelines.md
  + created   .amazonq/rules/agentsbridge.md
  + created   GEMINI.md

wrote 11 file(s)

Edit AGENTS.md, run agentsbridge sync, commit. That's the whole workflow.

Supported tools

Tool File generated Notes
Claude Code CLAUDE.md
GitHub Copilot .github/copilot-instructions.md
Cursor .cursor/rules/agentsbridge.mdc with alwaysApply frontmatter
Roo Code / Zoo Code .roo/rules/agentsbridge.md
Kilo Code .kilocode/rules/agentsbridge.md
JetBrains Junie .junie/guidelines.md
Amazon Q Developer .amazonq/rules/agentsbridge.md
Windsurf .windsurfrules
Cline .clinerules
Aider CONVENTIONS.md
Gemini CLI (legacy) GEMINI.md recent versions read AGENTS.md natively
Codex CLI, Gemini CLI, Jules, Amp, Zed, opencode — read AGENTS.md natively, nothing to generate

Run agentsbridge list to see this table in your terminal.

Missing a tool? Open an issue — adding a target is ~10 lines.

Safety

Generated files are marked:

<!-- generated by agentsbridge; do not edit -->

If a target file already exists without that marker (i.e. you wrote it by hand), sync skips it instead of clobbering your work. Use --force to adopt it deliberately.

agentsbridge sync --force        # adopt existing files
agentsbridge sync --dry-run      # preview changes
agentsbridge sync -t claude,copilot   # subset of targets

Tip: set AGENTSBRIDGE_TARGETS=claude,copilot in your environment (e.g. CI) to pin a subset without repeating -t on every invocation. Explicit -t always wins over the env var.

CI drift check

The whole point: rules that drift are rules nobody follows. Add the check to your workflow and CI fails whenever someone edits a generated file — or forgets to re-sync after editing AGENTS.md:

# .github/workflows/ci.yml
name: rules
on: [push, pull_request]
jobs:
  agentsbridge:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
      - uses: DavidStarYu/agentsbridge/action@main

Or invoke the CLI directly: agentsbridge check (exit code 1 on drift).

This repository runs the same check on itself — dogfooding.

FAQ

Why AGENTS.md as the source, not my own config? AGENTS.md is the emerging standard (adopted by OpenAI Codex, Gemini CLI, Jules, Amp, Zed and others — see agents.md). Tools that don't read it yet are exactly the ones this bridge targets. You keep zero agentsbridge-specific config: the source file is a standard.

Cursor already reads AGENTS.md, why generate an .mdc? Newer Cursor versions read AGENTS.md directly; many teams run versions that don't, or want explicit rule scoping. The generated .mdc is harmless if redundant — and you can exclude it: agentsbridge sync -t claude,copilot,....

Does it send my code anywhere? No. No network calls, no telemetry, no API keys. It's a template engine over one markdown file. Zero dependencies.

How is this different from X? rulesync (Node.js) covers more features — MCP servers, subagents, commands — at the cost of a bigger footprint and a .rulesync/ directory convention. agentsbridge does one thing: keep rule files in sync, in pure Python with zero dependencies. If you live in pipx/uv and want the simple thing, this is for you.

Development

git clone https://github.com/DavidStarYu/agentsbridge
cd agentsbridge
pip install -e . pytest ruff
pytest            # 47 tests
ruff check .

License

MIT

Metadata

Release files for agentsbridge 0.2.0

For a detailed explanation of source distributions (sdists) and built distributions (wheels), please see the package formats documentation.

Source distribution (sdist)

Source distribution for agentsbridge 0.2.0
File Size Uploaded
agentsbridge-0.2.0.tar.gz 15.0 kB Details

Built distribution (wheel)

Table of built distributions (wheels) for agentsbridge 0.2.0
File Interpreter ABI Platform
agentsbridge-0.2.0-py3-none-any.whl Python 3 none any Details

Total release size: 26.7 kB

Release files / agentsbridge-0.2.0.tar.gz

Download URL agentsbridge-0.2.0.tar.gz
Size 15.0 kB
Tags Source
SHA-256 checksum
How to use checksums
a282e9cc2415f88d503b75eaf79f9ee5ac39171f734065f71b5a024c551c5a7e
BLAKE2b-256 checksum
How to use checksums
5d60fbc1df0bf0930d116259ac1e5fb91948cbcfbd237054aea0bc9ac12ea4a0
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 3, 2026.

Transparency log

Release files / agentsbridge-0.2.0-py3-none-any.whl

Download URL agentsbridge-0.2.0-py3-none-any.whl
Size 11.7 kB
Tags Python 3
SHA-256 checksum
How to use checksums
174590257e9d7e2a16b9e527ade7b55223fe38b85b9cd24446e94c71f7f931a1
BLAKE2b-256 checksum
How to use checksums
aad160d916aa79f6f96e3405a151acaf972d8d57f6f3d0b0532eeb1702bdd5bf
Upload date
Uploaded using Trusted Publishing?
What is trusted publishing?
Yes
Uploaded via twine/7.0.0 CPython/3.13.14

Provenance

Provenance describes where a file came from. On PyPI, provenance is shared via attestations, which provide a verifiable record of the build or publishing details. View details, limitations and caveats.

PyPI Publish Attestation

PyPI verified that this artifact, at this checksum, originated from the publisher listed below.

Signed by GitHub Actions, verified by PyPI on Oct 3, 2026.

Transparency log

Release history Release notifications | RSS feed

This release

0.2.0 This release

2 release files

0.1.0

2 release files

Anthropic, PBC Visionary sponsor Bloomberg Visionary sponsor Hudson River Trading Visionary sponsor Meta Visionary sponsor NVIDIA Visionary sponsor Microsoft Sustainability sponsor Depot Continuous Integration AWS Cloud computing and Security Sponsor Datadog Monitoring Fastly CDN Google Download Analytics Sentry Error logging StatusPage Status page