ralphinho-rfc-pipeline

Breaks down large features into independently verifiable work units managed through a structured pipeline with quality gates, merge queues, and dependency tracking. Use this when a feature is too complex for a single agent pass and requires coordinated multi-unit implementation.

• Decomposes RFCs into a directed acyclic graph (DAG) of work units with explicit dependencies
• Assigns units to agents with scope, acceptance tests, risk levels, and rollback plans
• Enforces quality pipeline stages: research, planning, implementation, testing, review, and merge-ready reporting
• Manages merge queues with dependency validation and integration test re-runs after each merge
• Classifies work by complexity tier (isolated edits, multi-file changes, schema/auth/perf/security)
• Provides recovery workflows for stalled units via scope regeneration and constraint updates

1. 1.Write an RFC describing the feature or change
2. 2.Use the pipeline to decompose the RFC into a DAG of work units with the provided unit spec template (id, depends_on, scope, acceptance_tests, risk_level, rollback_plan)
3. 3.Assign each unit to an agent with its scope and constraints
4. 4.Execute the quality pipeline for each unit: research → planning → implementation → tests → review → merge-ready report
5. 5.Queue units for merge respecting dependency order; re-run integration tests after each merge
6. 6.Monitor the merge queue for dependency failures; evict stalled units and regenerate narrowed scope if needed
7. 7.Collect outputs: execution log, unit scorecards, dependency graph, and integration risk summary