cheddar

Pass

Audited by Gen Agent Trust Hub on Apr 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill relies on the membrane CLI to perform all operations, including authentication, connection setup, and action execution. These commands are executed in the local shell environment.
  • [EXTERNAL_DOWNLOADS]: The instructions require installing the @membranehq/cli package from the official NPM registry. This is a standard requirement for using the vendor's integration platform.
  • [PROMPT_INJECTION]: The skill has an indirect prompt injection surface because it interpolates user-provided queries (--intent "QUERY") and JSON inputs (--input '{"key": "value"}') directly into CLI commands.
  • Ingestion points: User-provided query strings and action parameters.
  • Boundary markers: None explicitly defined in the CLI examples.
  • Capability inventory: Ability to execute shell commands, manage network-connected actions, and handle authentication via the CLI.
  • Sanitization: Not explicitly mentioned in the documentation; safety depends on the agent's shell execution implementation.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 30, 2026, 03:20 PM
Security Audit — agent-trust-hub — cheddar