What problem does it solve?
Large AI audits miss files, features, and integrations because LLMs cannot reliably enumerate a repository, leaving security, product, and compliance questions open; power-mapper uses bash scripts and orchestrated agents to guarantee that every file is analyzed, every endpoint is traced, and every gap is documented before any executive report is issued.
Core Features & Use Cases
- Hierarchical MapReduce architecture: Tier 1 bash scripts capture every file, Tier 2 Sonnet agents analyze chunks, Tier 3 synthesizes domains, Tier 4 traces cross-cutting concerns, Tier 5 Opus generates FEATURES, GAPS, and AUDIT-SUMMARY, and Tier 6 verifies 100% coverage.
- Derivatives for downstream tooling: CODEBASE-CONTEXT.md energizes future AI sessions, DEPENDENCIES.md maps impact, SECURITY-BASELINE.md surfaces auth/API risks, TEST-MAP.md informs E2E planning, and CLEANUP.md lists dead-code targets.
- Use case: Run a full audit before onboarding a new team, planning a refactor, or delivering a security review so you have a deterministic feature inventory, integration audit, and technical debt report ready in minutes.
Quick Start
Ask power-mapper to run a full audit of this repository so every file, feature, and endpoint is mapped and verified.