datahub-connector-planning
Pass
Audited by Gen Agent Trust Hub on Mar 28, 2026
Risk Level: SAFE
Full Analysis
- [PROMPT_INJECTION]: The skill explicitly mitigates indirect prompt injection from research sources by instructing the agent to wrap external content in
<external-research>tags and ignore any embedded instructions, treating the results solely as data.- [COMMAND_EXECUTION]: Mandatory validation is performed on source system names to reject shell metacharacters and SQL syntax before the input is used in filesystem searches (find,grep) or package index queries (pip index).- [DATA_EXFILTRATION]: The skill encourages secure credential management by providing configuration templates that use environment variable placeholders (e.g.,${SOURCE_API_KEY}) rather than hardcoded secrets.- [EXTERNAL_DOWNLOADS]: The agent uses web research tools to fetch documentation and metadata specifications from well-known technology platforms and official registries to assist in the planning process.
Audit Metadata