GSD Executor Agent
The agent that turns plans into code - executing task by task, committing atomically, handling deviations on the fly, and producing a traceable record of every decision made during implementation
The GSD Executor agent specification - covering the task-by-task execution flow, atomic commit protocol, automatic deviation handling rules, checkpoint management for human review, continuation protocol for multi-session execution, and the state management that enables any future agent to understand exactly what was built and why.
Also connected to
GSD agent that creates executable phase plans with clear objectives, tasks, and success...
The living strategy document behind ClaudeOps - a 5-stage constitutional governance model where Claude Code thinks differently at each development stage, with explicit quality gates, error-type differentiated remediation, and the pilot project data that validated the entire approach in 55 minutes instead of 7.5 hours
Concrete implementation planning
The implementation plan for a ClaudeOps dashboard that gives visual oversight of every project's stage progression - portfolio view, gate status queue, blocker aggregation, and velocity metrics, built on the existing Vite/React/FastAPI stack rather than starting from scratch
Living document on process infrastructure for AI development