workable
Pass
Audited by Gen Agent Trust Hub on Sep 15, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted external data from Workable (such as candidate details, application contents, and comments). No explicit boundary markers or text sanitization protocols are defined in the instructions, which presents a surface where malicious external data could potentially influence downstream agent behavior during tool execution.
- [COMMAND_EXECUTION]: The instructions require the agent or user to execute several terminal commands using the
membraneCLI tool for authentication, connection management, and running data actions. - [EXTERNAL_DOWNLOADS]: The skill instructs the installation of the external global Node.js package
@membranehq/cli@latestto provide the runtime environment for interfacing with the platform.
Audit Metadata