pinterest-search
Pass
Audited by Gen Agent Trust Hub on Sep 11, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute the
postplusCLI tool using parameters such as--querythat are derived from user input as seen inSKILL.md. This represents a potential command injection surface if the agent does not properly sanitize the keyword input before shell execution. The skill also uses local file paths for output and challenge verification. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data from external Pinterest search results, which creates a vulnerability to indirect injection.
- Ingestion points: Search result metadata (titles, URLs) is ingested via the
postplus research runcommand. - Boundary markers: The skill does not define boundary markers or instructions to ignore commands within search results.
- Capability inventory: The agent is granted shell execution permissions for the
postplusCLI and file writing capabilities. - Sanitization: No sanitization or escaping of the external search data is specified in the instructions.
Audit Metadata