tools-and-context
Fail
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: CRITICALEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS] (LOW): The documentation recommends the installation of
@tambo-ai/reactand@modelcontextprotocol/sdk. These appear to be official SDKs for the Tambo platform and the standard Model Context Protocol. - [PROMPT_INJECTION] (LOW): Indirect Prompt Injection Surface. The skill enables the AI agent to process data from external sources and user-provided files.
- Ingestion points: Data enters the agent via
fetchUser,fetchDoc,fetchDocs, and theuseTamboContextAttachmenthook which processes local files. - Boundary markers: The provided examples do not demonstrate the use of delimiters or specific instructions to the LLM to ignore embedded commands in the context data.
- Capability inventory: Tools are used for data retrieval; the code examples do not include subprocess execution or sensitive file system modifications.
- Sanitization: The implementation uses
zodfor input and output schema validation, which provides structural integrity but does not sanitize against semantic instructions embedded within valid string fields.
Recommendations
- Contains 1 malicious URL(s) - DO NOT USE
Audit Metadata