Execute a complete iteration cycle: planning → batch execution → documentation → review setup. Follows file-based orchestration with TDD/ATDD, work logs, and quality gates.
Delegate task to specialist agent when expertise required: Initialize as specialist (Planning Petra, Backend Benny, Architect Alphonso, etc.) OR spawn sub-agent if available. Use when task requires specialized knowledge in agent's core focus area.
Test-first bug fixing: Write failing test that reproduces bug → Fix code → Test passes. Systematic, verifiable approach prevents regression and wastes no time on trial-and-error deployments.
Architect Alphonso conducts rigorous code review and architecture-fit analysis: ADR compliance, test coverage, architectural patterns, security. Outputs review document with APPROVED/REDIRECT/BLOCKED status.
Mid-execution self-monitoring checkpoint (Ralph Wiggum Loop): Detect warning signs (drift, confusion, gold-plating), course-correct before completion. Mandatory at 25% progress and before delegation.
Create functional specification for complex features: detailed requirements, user scenarios, constraints, acceptance criteria. Bridges strategic intent and implementation with MoSCoW prioritization and Given/When/Then scenarios.
Assess current implementation state: pending tasks, active work, progress metrics, blockers, and next recommended batch. Provides Planning Petra's executive view.
Instead of a centralized lookup table, each agent profile documents its own observed handoff patterns. This approach: - Preserves agent autonomy - Requires zero implementation complexity - Remains...