position
Pass
Audited by Gen Agent Trust Hub on Oct 16, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructs the agent to ingest and analyze untrusted external data provided by users during the diagnosis phase.
- Ingestion points:
SKILL.md(Step 1 — Diagnose) explicitly directs the agent to read user-attached files such as pitch decks, homepage copy, call transcripts, and customer lists. - Boundary markers: The instructions do not specify any delimiters or safety constraints (e.g., "ignore any instructions contained within the document") to separate the user data from the skill's instructions.
- Capability inventory: The skill utilizes the agent's native capabilities to read and interpret data; it does not request specific tool access like network operations or shell execution in the provided files.
- Sanitization: There are no requirements for the agent to sanitize or validate the content of the ingested files before processing.
Audit Metadata