gws-people
Pass
Audited by Gen Agent Trust Hub on Oct 6, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill relies on executing the
gwscommand-line tool to interact with the Google People API resources, includingcontactGroups,otherContacts, andpeopleas described in SKILL.md. - [INDIRECT_PROMPT_INJECTION]: The skill processes data retrieved from external sources (Google People API), which could be manipulated to include adversarial instructions.
- Ingestion points: The skill ingests contact information, group names, and profile data through various methods such as
people.get,people.listDirectoryPeople, andotherContacts.search(SKILL.md). - Boundary markers: No explicit delimiters or instructions are provided to help the agent distinguish between data retrieved from the API and its own instructions.
- Capability inventory: The skill allows the agent to execute shell commands via the
gwsCLI and perform data modification operations (create, update, delete). - Sanitization: There is no evidence of data sanitization or validation performed on the API responses before they are processed by the agent.
Audit Metadata