What problem does it solve?
Financial analysts, legal assistants, and researchers often struggle to turn long parsed documents into rigorous, auditable analysis. This Skill converts structured Markdown documents into four color-coded card types—blue (facts), green (explanations), yellow (risks), and red (actions)—with every claim traceable to a source fragment.
Core Features & Use Cases
- Four-Color Card Generation: Produces fact cards with source citations, explanation cards, risk cards flagging overstatements and uncertainty, and action cards with next-step recommendations.
- Multi-Stage Pipeline: Runs three stages (extract, review, propose) mapped to dedicated worker agents, with citation chaining enforced so green cards cite blue cards and red cards cite green/blue cards.
- Graceful Degradation: Falls back to a rule-based engine when the LLM is unavailable, marks low-confidence parses for manual review, and labels unreviewed risk cards as pending.
- Use Case: A financial analyst feeds a parsed research report into the skill and receives a structured JSON set of sourced facts, mechanism explanations, flagged overstatements, and recommended actions for an audit-ready report.
Quick Start
Run the four-color-cards skill on my parsed Markdown document to generate traceable fact, explanation, risk, and action cards.
