content-strategy
Pass
Audited by Gen Agent Trust Hub on Sep 14, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONNO_CODE
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze data from external sources including keyword research tools, customer call transcripts, survey responses, and web search results from forums like Reddit and Quora.
- Ingestion points: 'SKILL.md' describes processing keyword exports (Ahrefs, SEMrush), call transcripts, and external web content.
- Boundary markers: The instructions do not define specific delimiters for these external data sources or include directives to ignore potential instructions embedded within the analyzed content.
- Capability inventory: The skill's functionality is limited to text analysis and summarization. It lacks capabilities for arbitrary command execution, network exfiltration, or writing to sensitive system files.
- Sanitization: No explicit sanitization or validation of the external input is required by the instructions.
- [NO_CODE]: The skill consists entirely of high-level markdown instructions and does not include any accompanying scripts (Python, Node.js, Shell) or binary executables.
Audit Metadata