What problem does it solve? AI coding sessions lose all context when they end, forcing you to re-explain decisions, rediscover patterns, and repeat past mistakes across projects. This Skill persists decisions, error root causes, and architectural insights in a semantic memory graph that compounds across every session and project. ## Core Features & Use Cases - Recall Mode: Loads relevant past decisions and patterns at session start using project-prefixed semantic queries. - Capture Mode: Saves 2-5 richly-worded memories (decisions, errors, insights, preferences) after each completed task with tags and priorities. - Hypothesis Tracking: Records uncertain architectural decisions, collects supporting or weakening evidence, and verifies falsifiable predictions. - Maintenance Mode: Runs health checks, consolidates episodic memories into semantic ones, and prunes outdated or orphaned entries. - Use Case: After debugging a stale-closure bug in a React project, the Skill saves the root cause and fix pattern; months later, a different project with a similar symptom automatically surfaces that memory during recall. ## Quick Start Ask the agent to recall what it remembers about your current project's authentication decisions before starting work.
