marketing-social
Pass
Audited by Gen Agent Trust Hub on Aug 3, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill requires the installation of the agent-browser tool from a trusted source (github.com/vercel-labs/agent-browser). This is a standard dependency for the skill's research functionality.
- [COMMAND_EXECUTION]: The skill provides clear instructions and shell commands for the agent to perform automated browsing tasks, such as extracting text and taking screenshots from competitor social media profiles. These operations are scoped to the intended marketing research purpose.
- [PROMPT_INJECTION]: The skill ingests external data from social media websites (Instagram, TikTok, LinkedIn) to inform its strategies. While this constitutes a surface for indirect prompt injection, it is a functional requirement for competitive analysis, and no evidence of exploitation was found. The agent relies on internal safety guardrails when processing this external content.
Audit Metadata