paw-mkt-influencer
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the
agent-browsertool and shell scripts for influencer discovery, metrics extraction, and environment setup across several reference files. - [EXTERNAL_DOWNLOADS]: Recommends downloading and installing the
agent-browserutility from the Vercel Labs GitHub repository, which is a well-known and trusted source. - [INDIRECT_PROMPT_INJECTION]: The skill has an inherent vulnerability surface due to its primary research function. 1. Ingestion points: Content is ingested from TikTok, Instagram, and YouTube search results as documented in influencer-identification.md. 2. Boundary markers: The instructions focus the agent on specific metric extraction (e.g., follower counts and engagement rates) which helps scope the data processing. 3. Capability inventory: The skill utilizes browser automation and local file system operations to generate reports. 4. Sanitization: No explicit content sanitization protocols are defined beyond the instruction to extract specific data fields.
Audit Metadata