save-pattern

What problem does it solve?

This Skill provides a centralized way to store reusable application patterns in AgentDB's patterns table, enabling consistent knowledge sharing across agents through semantic embeddings.

Core Features & Use Cases

• Store architectural and procedural patterns with semantic embeddings for efficient search via agentdb_pattern_search.
• Prevent duplicates by first searching existing patterns before storage, ensuring high-quality pattern maintenance.
• Retrieve pattern usage statistics with agentdb_pattern_stats to monitor adoption and quality.

Quick Start

Use the save-pattern skill to store a new pattern by calling mcp__agentdb__agentdb_pattern_store with taskType, approach, and optional tags and successRate. Example: mcp__agentdb__agentdb_pattern_store( taskType="architecture:domain-adapter", approach="Domain Adapter Pattern for Data Sources: All data sources implement the Source trait, fetch() -> Vec<TimeSeriesPoint>, and health_check() -> HealthStatus. Related files: core/src/traits.rs, core/src/sources/http_poll.rs", successRate=0.95, tags=["architecture","pattern","agentdb"] )