trackline mcp: a check-before-you-act server for coding agents
Economy_Rice_8862 · reddit · 2026-09-27
A dev shares trackline mcp, exposing trackline's rule checking to any MCP client so coding agents can self-check before acting.
Two tools:
- getrules: returns project rules so the agent holds them in context;
- checkaction: the agent describes its intended action and gets a verdict with reasoning.
Positioning and limits are stated plainly: in Claude Code, Codex and Cursor, trackline runs as a hook and can actually block; over MCP the agent decides whether to ask — no checkaction call means no oversight — so it's advisory, not enforcement, extending reach to hook-less agents rather than replacing hooks. Runs locally over stdio, no account or API key: npm install -g trackline, then trackline mcp. The author admits the weak point is getting agents to consistently call a check-first tool and asks the community for approaches.
More from coding & agent
- Adobe MAX: MCP across Creative Cloud, AI agents editing layers, Firefly Builder and Graph — umesh_ai · 2026-11-11
- Sydney Codex Community Buildathon set for Nov 14 with Cerebras-backed ultrafast compute track — paw_lean · 2026-11-11
- shadcn's improve hits 10k stars: let top models plan, cheap models execute code fixes — shadcn · 2026-11-11
- OpenAI open-sources Codex Security CLI and SDK for AI-driven vulnerability scanning and fixing — openai · 2026-11-11
- Claude Opus 5.5 runs an architecture search and finds AttnRes-Canon, a new top Transformer design — SonglinYang4 · 2026-11-11
- 5 research habits worth knowing: read papers backwards, reproduce first, wire coding agents via MCP — cneuralnetwork · 2026-11-11