clyde
Pass
Audited by Gen Agent Trust Hub on Apr 22, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill instructs the user to install the
@membranehq/clipackage globally via npm. This is a legitimate tool provided by the skill vendor for interacting with their platform. - [COMMAND_EXECUTION]: The skill relies on executing shell commands via the
membraneCLI to authenticate, manage connections, and perform API requests. This is the intended operation of the skill. - [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface because it processes data retrieved from the external Clyde API, which could contain instructions designed to influence the agent.
- Ingestion points: Data returned from
membrane action runandmembrane requestcommands is ingested into the agent's context. - Boundary markers: None are specified in the instructions to separate external data from system instructions.
- Capability inventory: The skill uses subprocess execution of the
membraneCLI tool. - Sanitization: There is no evidence of sanitization or validation performed on the data received from the Clyde API before it is processed by the agent.
Audit Metadata