What problem does it solve?
The programmer Markov brain automates and governs implementation work on an Android project by reading PM context, selecting backlog items, producing focused plans, making minimal code changes, running tests and health checks, and recording results for future sessions. It reduces context-switching, prevents scope creep, and enforces project conventions so engineers can deliver reliable changes with clear handoffs.
Core Features & Use Cases
- Backlog-driven implementation: Picks highest-priority OPEN items from docs/backlog.md and marks progress.
- Context-first edits: Reads docs/programmer-brief.md and per-package CLAUDE.md before changing code.
- Iterative loop: Plan -> Write code -> Write tests -> Run tests -> Run health check -> Record changes -> Update backlog.
- Automated decision tools: Uses scripts to classify state, recommend next phase, detect scope creep, and log transitions.
- Use Case: Triage a SEV-1 bug, implement the minimal fix, run unit tests and project health check, update CLAUDE.md, and mark the backlog item DONE with evidence.
Quick Start
Implement the highest-priority OPEN backlog item by reading the PM brief and package CLAUDE.md, plan a minimal change, write the code and tests, run tests and the health check, then update CLAUDE.md and docs/backlog.md.