The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill requires the installation of the agentdb and agentic-flow packages from the npm registry to provide database and reasoning capabilities.
[COMMAND_EXECUTION]: Provides instructions to initialize local databases, migrate legacy data, and configure an MCP server using npx agentdb. These commands are standard for the described functionality.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process agent 'trajectories' and 'experiences'. This creates a surface for indirect prompt injection if the stored historical data includes adversarial instructions that are later retrieved and synthesized into the agent's context.
Ingestion points: Reads patterns and memories from local database files (.agentdb/reasoningbank.db and .swarm/memory.db) as shown in SKILL.md.
Boundary markers: None identified in the provided TypeScript examples.
Capability inventory: Accesses the file system for database management and executes shell commands for MCP integration as described in SKILL.md.
Sanitization: Relies on standard JSON serialization for pattern data storage.

reasoningbank-agentdb