ad-account-diagnostic
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted external data in the form of ad platform exports (campaign, ad-set, and ad-level data) and CRM/backend revenue records to perform its diagnostics.
- Ingestion points: Untrusted data enters the agent context through the "exports" and "CRM or backend revenue data" requested in the "Interview" and "Workflow" sections of SKILL.md.
- Boundary markers: The skill partially mitigates risk by requiring an interview to be completed before any data is processed, but it does not specify explicit delimiters (like XML tags or triple backticks) for the ingested data itself.
- Capability inventory: The skill performs complex metric decomposition, generates spreadsheet formulas, and utilizes persistent memory (if available in the host environment) to store account baselines and verdicts.
- Sanitization: There is no mention of escaping, validation, or filtering of the content within the ad exports or CRM data before it is processed by the agent.
Audit Metadata