phronesis — AI-native Advisory Practice

Status: Phase 1 complete (v0.1.2-pre) — substrate + 15 active frameworks + 5 stages + 7 deliverables + L1–L5 linter + M7 extraction pipeline. Phase 2 engagement-driven (M4 stage CLI + M5 autonomous-mode stub-to-real path). Track progress: docs/methodology.md (canonical reference), docs/adr/0001-0008, and the design spec at ~/broomva/docs/superpowers/specs/2026-05-06-phronesis-design.md.

What it does

Runs AI/digital consulting engagements through 5 stages:

1. Intake — stakeholder map, interview guide, strategic thesis (kills "hagamos algo de IA" — Bision empirical 100% failure)
2. Maturity Scan — MIT CISR + Gartner AI dimensions, gap analysis, capability heatmap
3. Use-Case Ideation — JTBD + VPC, with IdeationSource diversity check
4. Prioritization — RICE/ICE/WSJF, ROI Model with sensitivity, Impact-Effort Matrix
5. Roadmap Synthesis — Three Horizons, QuantumBlack ML lifecycle pilots, baselines captured

Each stage has a non-negotiable human-review gate.

When to invoke this skill