memory-ingest
Pass
Audited by Gen Agent Trust Hub on Oct 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process unstructured external content (meeting transcripts, conversation logs, pasted documents) which is a primary vector for indirect prompt injection.
- Ingestion points: Step 1 (Parse Raw Input) and Step 2 (Extract Entities) read and process raw external text provided by the user or from external files.
- Boundary markers: The skill does not define or instruct the agent to use specific boundary markers or delimiters when processing untrusted input to separate data from instructions.
- Capability inventory: The agent has access to
search_notesandwrite_notecapabilities to modify the knowledge base based on the ingested content. - Sanitization: The instructions explicitly tell the agent to preserve source content verbatim in the created notes without mentioning sanitization, validation, or escaping of potentially malicious instructions embedded in the raw text.
Audit Metadata