clay
Pass
Audited by Gen Agent Trust Hub on Aug 24, 2026
Risk Level: SAFEPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes data from an external source (Clay tables).
- Ingestion points: Data is ingested from the Clay API in
scripts/listRecords.tsandscripts/findRecord.ts. - Boundary markers: The skill does not implement specific boundary markers or instructions to delimit external data from agent instructions in its script outputs.
- Capability inventory: The skill provides write capabilities via
createRecord.tsandupdateRecord.ts, and navigation tools that could be influenced by malicious data. - Sanitization: No evidence of automated filtering or escaping is present for content retrieved from the Clay API.
- [EXTERNAL_DOWNLOADS]: The skill documentation recommends using
npxto run or install the connector from@zapier/clay-connector. These references target a well-known package registry and a trusted organization, representing standard deployment behavior for this vendor.
Audit Metadata