denney-persona-angle-format
Pass
Audited by Gen Agent Trust Hub on Oct 8, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill provides a strategic framework for creative creative strategy. It consists entirely of markdown instructions and templates, without any executable code, shell commands, or tool definitions that could be exploited for unauthorized access or behavior.- [INDIRECT_PROMPT_INJECTION]: The skill involves analyzing untrusted customer reviews and social media comments (SKILL.md, Procedure step 2). This risk is assessed as safe because:
- Ingestion points: Untrusted text enters the context through manual analysis tasks.
- Boundary markers: The skill requires using locators to trace evidence and separating observed facts from AI interpretations.
- Capability inventory: The skill has no dangerous capabilities, as no tools, network access, or file-writing operations are defined.
- Sanitization: The instructions explicitly mandate that the agent's conclusions must be verified against original source materials and that the strategist should not rely blindly on AI-organized patterns.
Audit Metadata