marketing-paid-ads
Pass
Audited by Gen Agent Trust Hub on Apr 27, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill utilizes the agent-browser tool from a trusted organization for its automated research tasks. The dependency is installed using standard package management commands as outlined in the configuration and shared patterns.
- [COMMAND_EXECUTION]: The skill executes shell commands to run browser automation for tasks such as capturing text and screenshots from various ad platforms and competitor sites. These commands are necessary for the competitive intelligence functionality.
- [PROMPT_INJECTION]: The skill's research mode exposes it to indirect prompt injection. 1. Ingestion points: External text is ingested from competitor landing pages and third-party ad libraries via browser automation in SKILL.md. 2. Boundary markers: The skill does not employ delimiters or specific instructions to the agent to disregard commands within the scraped text. 3. Capability inventory: The skill environment allows reading brand strategy files and performing further network actions through the browser. 4. Sanitization: There is no evidence of sanitization or validation of the data retrieved from external URLs before it is processed by the agent.
Audit Metadata