run-product-review

Pass

Audited by Gen Agent Trust Hub on Oct 27, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process potentially untrusted external data, which creates a surface for indirect prompt injection attacks.\n
  • Ingestion points: The workflow in SKILL.md involves collecting 'customer evidence' and 'surveys'. Additionally, the 'BuildBetter Acceleration' section mentions processing 'representative evidence', 'linked tickets', and 'documents'.\n
  • Boundary markers: The instructions do not define explicit delimiters or boundary markers to help the agent distinguish between its system instructions and the content of the ingested data.\n
  • Capability inventory: The agent aggregates this data to generate summaries, agendas, and decision queues. While no direct command execution is present in the provided files, the output influences the agent's decision-making context.\n
  • Sanitization: There are no instructions for the agent to sanitize, escape, or validate the content of external data before processing it.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 27, 2026, 03:14 AM