google-local-services-search
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONDYNAMIC_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill processes external search results from the Google Local Services API, which represents an untrusted data surface.
- Ingestion points: External content is retrieved from the
/service/google-local-servicesendpoint inscripts/search_google_local_services.sh. - Capability inventory: The skill can write output to a user-defined local file path and prints data to stdout.
- Boundary markers: None present in the instructions.
- Sanitization: The script includes a validation step that ensures the API response is valid JSON before processing.
- [DYNAMIC_EXECUTION]: The skill uses inline Python scripts to handle complex input validation and JSON parsing tasks.
- Evidence: Fixed Python templates are embedded within
scripts/search_google_local_services.shand executed at runtime to process arguments and validate the API response. - [COMMAND_EXECUTION]: The skill utilizes standard system utilities to perform its operations.
- Evidence: It uses
curlto interact with the vendor's API and standard shell commands likemkdirandcpto manage output files when the--outputflag is used.
Audit Metadata