RuleDrift
Test whether AI agents retain critical instructions as conversations grow.
Open source Open in the app JSON README (API)
About
Test whether AI agents retain critical instructions as conversations grow.
Details
- Kind
- MCP servers
- Topic
- No topic detected
- Publisher
- ri7in
- Origin
- official
- Category
- ferramentas
- Transport
- local
- Version
- 0.1.1
- Stars
- 1
- Open pull requests
- 5
- Last push
- 2026-09-03T04:05:35Z
- Repository state
- ativo
- Language
- TypeScript
- License
- MIT
- Added
- 2026-08-29 04:01:17
- Updated
- 2026-08-29 04:01:17
- Origin id
io.github.ri7in/rule-drift
README
<div align="center"> <img src="docs/assets/rule-drift-mark.svg" width="88" height="88" alt="RuleDrift logo"> <h1>RuleDrift</h1> <p><strong>Catch the moment your agent drifts from its rules.</strong></p> </div> RuleDrift tests whether an AI agent still follows important instructions after a conversation becomes long or distracting. Run it like a test suite when you change a prompt, model, memory system, or agent workflow. It runs locally and does not sit in front of production requests. ## Quick start Requires Node.js 22+ and [Ollama](https://ollama.com/). ```bash ollama pull qwen3:1.7b npx rule-drift init npx rule-drift test ``` No paid model API, account, backend, or telemetry is required. ## What you get - The first conversation point where a rule fails - Terminal, JSON, and JUnit reports - Local Ollama and bring-your-own-agent support - Optional MCP tools for AI coding clients - Stable exit codes for CI ## MCP ```bash npx -y rule-drift mcp --root /path/to/your/project ``` See the [MCP setup guide](docs/mcp.md) for client configuration and safety details. ## Documentation - [Full guide](docs/guide.md) - [All documentation](docs/README.md) Contributions are welcome—see [CONTRIBUTING.md](CONTRIBUTING.md). Licensed under [MIT](LICENSE).