What problem does it solve?
Maestro Analyze automates multi-dimensional technical analysis that would otherwise require coordinating many human reviewers and manual note-taking. It converts exploration, evidence gathering, and multi-axis scoring into a reproducible CSV-driven pipeline so teams can get objective, auditable Go/No-Go recommendations and decision context quickly.
Core Features & Use Cases
- Wave-based multi-agent orchestration: runs exploration agents, six-dimension scoring agents, and a single synthesis agent in a diamond topology to produce a consolidated verdict.
- CSV-first state and append-only discovery board: uses tasks.csv and discoveries.ndjson as the single source of truth for provenance, retries, and merging agent outputs.
- Quick and full modes: supports a quick synthesis-only shortcut or a full exploration → score → synthesize flow with concurrency controls and graceful degradation.
- Use case: evaluate a proposed Phase 3 feature by extracting code anchors, scoring feasibility/impact/risk/complexity/alignment/maintainability, and generating analysis.md, context.md, and conclusions.json for planners and implementers.
Quick Start
Run the maestro-analyze command with a subject like "microservices vs monolith" (optionally -y to auto-run) to produce an analysis.md, context.md, and the CSV session artifacts in .workflow/.csv-wave.
