do-marketing
Pass
Audited by Gen Agent Trust Hub on Oct 7, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect prompt injection by ingesting and processing external data such as product documentation, search engine results, and AI citations. While the instructions emphasize data privacy and redaction, they do not include explicit instructions for the agent to disregard instructions potentially embedded within these external sources.
- Ingestion points: Reading product surfaces, documentation, analytics code, and repository history as specified in SKILL.md; evaluating search engine results and AI citations as described in reference/playbooks.md.
- Boundary markers: No specific delimiters or 'ignore embedded instructions' warnings are present to mitigate adversarial content within the ingested data.
- Capability inventory: The skill generates content such as reports, decks, scripts, and video storyboards, and can interact with other installed skills for design and art creation.
- Sanitization: Includes instructions to redact private evidence and exclude financial data from all outputs.
Audit Metadata