outreach-manager
Pass
Audited by Gen Agent Trust Hub on Jul 6, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill is entirely instruction-based and focuses on generating text for marketing campaigns. It does not perform network operations or install any external packages.
- [DATA_EXFILTRATION]: Information management is confined to the local memory directory (e.g., memory/creators/, memory/influencer/) to persist partnership status. No sensitive files like credentials, SSH keys, or AWS configurations are accessed.
- [COMMAND_EXECUTION]: The instructions and templates contain no shell commands, subprocess calls, or instances of dynamic context injection syntax.
- [REMOTE_CODE_EXECUTION]: No remote scripts are downloaded or executed. The skill relies on Tier 1 data sources, which means all information is either provided by the user or retrieved from local memory files.
- [PROMPT_INJECTION]: The skill processes external data (influencer handles and campaign details) to generate copy, but it lacks the automated capabilities (like code execution or web access) required to facilitate a dangerous indirect injection attack. No instructions designed to bypass safety filters or extract system prompts were detected.
Audit Metadata