What problem does it solve? Decisions made during a coding session are usually lost or scattered across chat logs, making it impossible to trace why a choice was made or find prior precedents. This Skill structures that reasoning into a queryable knowledge graph with provenance, then bridges it into persistent memory so insights survive the session. ## Core Features & Use Cases - Decision Recording with Provenance: Record decisions with category, scenario, reasoning, outcome, and confidence, then cite decision IDs in commits or PRs for traceability. - Entity Extraction & Causal Tracing: Extract entities and relationships from conversations via agent-guided reasoning, and trace causal chains or find precedents with graph queries. - Automatic Graph Persistence: PostToolUse hooks auto-save every export_graph result to per-session timestamped files under .semantica/, which AiRaccoon memory watches and ingests for cross-session retrieval. - Use Case: During an architecture review, record each design decision with its rationale, link the affected components as graph entities, and later run find_precedents to check whether a similar scenario was already decided. ## Quick Start Ask the agent to record the current design decision in the Semantica knowledge graph with its reasoning and confidence, then query the causal chain behind it.
