# RuleMetric

> RuleMetric measures whether changes to AI coding-agent instructions — CLAUDE.md, AGENTS.md, rules, memories and skills — improve results. It captures sessions from Claude Code, Codex, Cursor and GitHub Copilot, surfaces suggestions, and runs controlled comparisons that report quality, cost and latency with sample sizes and uncertainty.

These pages are plain markdown so agents can read them without running JavaScript. The human docs at /docs cover the same material.

## Docs

- [Project status and daily review](https://rulemetric.com/agent-docs/project-status.md): Daily source and GitHub review, pending work, checked delivery state and documentation upkeep.
- [Quick start](https://rulemetric.com/agent-docs/quickstart.md): Install the CLI, authenticate, capture sessions from each supported tool, run the worker, verify setup, and Local Mode.
- [Integrating an agent](https://rulemetric.com/agent-docs/integrate.md): How an AI agent should set up RuleMetric and use the CLI, HTTP API or MCP server; auth, errors, and how to read results.
- [MCP tools and resources](https://rulemetric.com/agent-docs/mcp-tools.md): Generated from real MCP discovery: input fields, structured output availability, read-only declarations and documentation resources.
- [Suggestions page guide](https://rulemetric.com/agent-docs/suggestions.md): Review actions, scores, testing, retirement, dismissal, adoption history and reading current proposal data.
- [CLI reference](https://rulemetric.com/agent-docs/cli.md): Every `rulemetric` command, argument and flag (generated from @rulemetric/cli 0.26.0).
- [HTTP API reference](https://rulemetric.com/agent-docs/api.md): Authentication and the REST endpoints for sessions, instructions, evaluations and insights.

## Optional

- [Everything in one file](https://rulemetric.com/llms-full.txt): all of the pages above concatenated
- [Terms of Service](https://rulemetric.com/terms)
- [Privacy Policy](https://rulemetric.com/privacy)