What problem does it solve?
Understanding past AI decisions, finding relevant context, identifying patterns, and debugging the decision graph memory system can be complex and time-consuming. This skill provides tools and patterns to effectively query, analyze, and troubleshoot the AI Counsel's institutional memory.
Core Features & Use Cases
- Similar Decision Retrieval: Find and inject context from past deliberations semantically related to a new question.
- Database Statistics & Health: Monitor the growth, check the integrity, and debug performance of the SQLite-based decision graph.
- Cache Performance Analysis: Analyze L1/L2 cache hit rates to optimize query latency and memory usage.
- Tiered Context Formatting: Understand how context is formatted (strong, moderate, brief) based on similarity and token budget.
- Use Case: Before starting a new deliberation on "Should we use GraphQL?", use this skill to query the decision graph for similar past debates, injecting relevant context to accelerate convergence and avoid re-hashing old arguments.
Quick Start
To find past deliberations related to a specific topic, use the query_decisions tool with the search_similar operation. For example:
mcp__ai-counsel__query_decisions({
query: "database choice",
operation: "search_similar",
limit: 5
})
