content-strategy
Pass
Audited by Gen Agent Trust Hub on Jun 24, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns such as credential harvesting, obfuscation, or persistence mechanisms were detected in the skill instructions or metadata.- [SAFE]: The skill includes instructions to process external data (Category 8: Indirect Prompt Injection) such as customer call transcripts, forum posts from Reddit/Quora, and competitor blogs. This is considered safe as the skill lacks any high-risk capabilities to exploit malicious instructions. \n
- Ingestion points: SKILL.md (sections for Keyword Data, Call Transcripts, Survey Responses, Forum Research, and Competitor Analysis). \n
- Boundary markers: Absent; the instructions do not suggest using delimiters to wrap external data. \n
- Capability inventory: No dangerous tools, subprocess calls, file-write operations, or dynamic code execution were found in the skill. \n
- Sanitization: Absent; the agent is expected to read and summarize the content directly for planning purposes.
Audit Metadata