clay-skill-author
Pass
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill includes a self-update mechanism that checks for newer versions at the official GitHub repository of the vendor (
clay-run/clay-skill-creator). If an update is detected, it downloads the current skill bundle to ensure the user is running the latest logic and security rules. All update sources are restricted to the verified vendor's own domains. - [COMMAND_EXECUTION]: The tool uses
subprocess.runto interact with the localclayCLI and its own validation scripts. These operations are strictly limited to an allowlist of read-only configuration commands required to extract the structure of the user's workflows or tables. This functionality is core to the skill's purpose as a conversion and authoring tool. - [SAFE]: The skill is designed with a strong focus on security and data privacy. It includes dedicated modules (
scripts/injection.pyandscripts/portability.py) used to detect prompt injection and hardcoded credentials in the skills it generates. The instructions emphasize that the tool only reads configuration metadata and never accesses sensitive data rows or exfiltrates user credentials.
Audit Metadata