Ruler
One source of truth for AI coding-agent instructions — apply the same rules to every agent.
About
Ruler solves the config sprawl that appears the moment a team uses more than one AI coding tool: Claude Code wants CLAUDE.md, Codex wants AGENTS.md, Copilot and Cursor each have their own files, and keeping them consistent by hand is a losing game. You write your rules once as Markdown files in a .ruler/ directory, declare which agents to target in ruler.toml, and ruler apply distributes everything to the right config files.
It also propagates MCP server settings across agents, keeps generated files out of version control via .gitignore automation, supports nested .ruler/ directories for context-specific rules in monorepos, and ships as a simple CLI on npm. The project is MIT-licensed, actively maintained (commits within the last week), and carries a beta-research-preview banner — test it in your environment before standardizing on it.
What it can do
- Centralized rule management in a `.ruler/` Markdown directory
- Automatic distribution to every configured agent's instruction file
- Per-agent targeting and output paths via `ruler.toml`
- MCP server settings propagated across agents
- Nested rule loading for context-specific instructions
- .gitignore automation for generated configs; `ruler revert` for clean undo
Setup
- Install with
npm install -g @intellectronica/ruler. - Create a
.ruler/directory and write your rules as Markdown files (zero-padded numeric prefixes control ordering). - Declare target agents and output paths in
ruler.toml. - Run
ruler apply(use--dry-run --verboseto preview);ruler revertundoes generated files.
Prerequisites
Node.js for the npm install; at least one supported AI coding agent (Claude Code, Codex, Copilot, Cursor, Aider, and others).
Use cases
Works with
Trust score — 7/10
Scored 7/10 (metrics tier, metrics-v1). Evidence: 2,945 GitHub stars (GitHub API), MIT license, pushed Oct 14 2026, README feature claims corroborated by real community adoption (repos managing conventions through it). Strengths: open, small focused tool, honest beta labeling. Docked for: (1) modest adoption vs. the problem's size; (2) beta research preview — expect rough edges; (3) generated-file workflows can surprise teams that don't read the .gitignore behavior. License: MIT.
★ 2,945 GitHub stars
How the trust score is built
Full methodology →- Adoption · stars/downloads, log-scaled vs peers
- 6.0/10
- Maintenance · recency, releases, issue load
- 8.0/10
- Openness · license + docs
- 10.0/10
- Manual review · sanity read + red-flag check
- 7.0/10
- · 2,945 GitHub stars (log-scaled vs catalog peers, verified via GitHub API)
- · last push Oct 14 2026; MIT-licensed; npm @intellectronica/ruler
- · Single .ruler/ source of truth distributed to CLAUDE.md/AGENTS.md/etc, MCP server propagation, .gitignore automation — README claims corroborated by community repos managing conventions through it
- · manual review: honest beta-research-preview labeling; modest adoption vs problem size; generated-file workflows can surprise teams that skip the .gitignore behavior
Sources
Every fact above was checked against these sources. Where a claim comes from the provider and couldn't be independently verified, the text says so.