audiences

Pass

Audited by Gen Agent Trust Hub on Aug 15, 2026

Risk Level: SAFENO_CODE
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill facilitates the ingestion of external data from a workspace's CRM records (people, companies, and deals) which could potentially contain instructions intended to influence the agent's behavior.
  • Ingestion points: Data is brought into the agent's context through the clay audiences records get and clay audiences records search-ids commands, which retrieve field values for records stored in the workspace.
  • Boundary markers: The skill does not explicitly instruct the agent to use delimiters or ignore embedded instructions within individual record field values.
  • Capability inventory: The skill relies on the clay CLI and jq for data processing, as well as file writing to the /tmp/ directory for cataloging fields.
  • Sanitization: No specific sanitization or filtering of the retrieved record content is mentioned in the instructions.
  • [NO_CODE]: The skill does not contain any executable scripts, binaries, or hidden code. It provides documentation and CLI usage instructions, relying entirely on the pre-existing environment and the 'clay' vendor utility.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 15, 2026, 11:01 PM