pinterest-search
Pass
Audited by Gen Agent Trust Hub on Jul 20, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the 'postplus' CLI utility for various operations, including searching and diagnostics. These involve running shell commands that process JSON-formatted request and result data.
- [PROMPT_INJECTION]: The skill is exposed to indirect prompt injection risks through the ingestion of external data from Pinterest.
- Ingestion points: External data enters the agent context via 'result.json', which contains pin titles and URLs retrieved from Pinterest.
- Boundary markers: Absent; no delimiters or instructions are provided to distinguish external search content from trusted instructions.
- Capability inventory: The skill possesses capabilities for shell command execution and file system access, which could be leveraged if malicious instructions in the search results were followed.
- Sanitization: Absent; there are no specified procedures for sanitizing or validating the content of the search results.
Audit Metadata