What problem does it solve? Long-term investors struggle to distinguish real fundamental changes from mere rewording or price swings when reviewing their holdings. This Skill compares two versions of an investment thesis report and determines whether the underlying facts actually changed, preventing both overreaction to noise and blindness to genuine deterioration. ## Core Features & Use Cases - Evidence-Based Drift Detection: Compares valuation anchors, core assumptions, red lines, management quality, and competitive moat across two report snapshots, classifying each dimension as Improved, Unchanged, or Weakened. - Three Operating Modes: Supports explicit two-report comparison, automatic snapshot discovery in the reports directory, and graceful handling when no historical baseline exists. - Verified Numerics: Routes all valuation, percentage, and market-cap calculations through tools/financial_rigor.py to eliminate LLM arithmetic errors, with cross-source validation requirements. - Use Case: After a new earnings release, compare your original Pinduoduo thesis snapshot against an updated report to learn whether margin compression is a real thesis break or just a price-driven narrative shift, and get a migrated action recommendation (Buy/Hold/Reduce/Exit). ## Quick Start Ask the AI to run a thesis drift check on a company by providing its name along with the paths to the old and new thesis reports, or just the company name to auto-discover snapshots in the reports folder.
