marketing-strategy-pmm
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONMETADATA_POISONING
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill instructions and templates involve processing data from untrusted external sources, creating a potential surface for indirect prompt injection attacks.\n
- Ingestion points: The skill guides the agent to process customer interview notes, Gong/Chorus sales call transcripts, and public reviews from platforms like G2 or Capterra.\n
- Boundary markers: The provided templates lack explicit delimiters or 'ignore embedded instructions' warnings for handling this external text.\n
- Capability inventory: The skill references Python scripts and tool integrations (HubSpot, Productboard) that may perform actions or store data based on the content of analyzed inputs.\n
- Sanitization: No sanitization or validation logic is defined for the ingestion of these external inputs.\n- [METADATA_POISONING]: The skill contains a discrepancy in its metadata, listing 'Alireza Rezvani' as the author in the YAML frontmatter, which does not match the provided author context of 'davila7'.
Audit Metadata