content-strategy
Pass
Audited by Gen Agent Trust Hub on Nov 1, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from sales transcripts, survey responses, and forum search results. 1. Ingestion points: SKILL.md ideation sections. 2. Boundary markers: Absent. 3. Capability inventory: Natural language analysis only; no scripts or network tools. 4. Sanitization: Absent.
- [DATA_EXFILTRATION]: The skill accesses local marketing context files (e.g., .agents/product-marketing.md) to ground its analysis. This is standard functional behavior for the agent.
Audit Metadata