sales-outreach-personalization
Pass
Audited by Gen Agent Trust Hub on Oct 26, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to process data from external, potentially untrusted sources such as community forums, review platforms, and general web search results (referenced in 'SKILL.md' and 'references/signal-taxonomy.md'). This establishes an attack surface for indirect prompt injection, where malicious instructions embedded in analyzed prospect data could influence agent behavior during hook generation. The workflow lacks explicit instructions for using boundary markers or delimiters to isolate interpolated external data within the prompt context. While the skill includes an 'Anti-fabrication rule' and a 'Self-check gate' to verify signal accuracy, these focus on factual integrity rather than technical sanitization or protection against adversarial instructions in the input data.
Audit Metadata