ReasoningBank with AgentDB
Pass
Audited by Gen Agent Trust Hub on Mar 1, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documentation suggests using
npx agentdb@latestto initialize the database and run the MCP server. This command fetches the latest version of the agentdb package from the NPM registry. - [COMMAND_EXECUTION]: Provides CLI commands for database initialization, data migration, and statistics reporting. These operations interact with local database files such as
.agentdb$reasoningbank.dband.swarm$memory.db. - [PROMPT_INJECTION]: Features a surface for indirect prompt injection as it is designed to ingest and store external data (trajectories and experiences) which are then retrieved for agent reasoning.
- Ingestion points: Data is ingested via the
insertPatternmethod and theimportCLI command described inSKILL.md. - Boundary markers: Not explicitly defined in the provided code snippets; the skill relies on the underlying LLM's interpretation of stored patterns.
- Capability inventory: The skill facilitates file system writes for database storage and provides CLI tool access for management.
- Sanitization: No specific sanitization or filtering of the ingested pattern data is demonstrated in the examples.
Audit Metadata