What problem does it solve? Project context scattered across multiple .ai/ state files (BRIEF, TASKS, LOG, REVIEWS) becomes stale and hard to consume, leaving agents and developers without a current, compact summary of project health. ## Core Features & Use Cases - Multi-file aggregation: Reads .ai/BRIEF.md, TASKS.md, architect.md, REVIEWS.md, LOG.md, and .mcp.json to build a unified snapshot. - Structured 20-60 line digest: Produces a bullets-only DIGEST.md with fixed sections covering product, stack, triad health, current focus, decisions, risks, MCP servers, and recent changes. - Staleness triggers: Runs when DIGEST.md is over 3 days old, after a major sprint, or after running the ai-archive skill. - Use Case: After a sprint where the engineer closed several E-## tasks and critics added findings to REVIEWS.md, run this Skill to regenerate DIGEST.md so the next session starts with an accurate project snapshot. ## Quick Start Ask the AI to regenerate the project digest because DIGEST.md is stale after the latest sprint.