fidel

Pass

Audited by Gen Agent Trust Hub on Apr 21, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill's primary functionality is delivered through the execution of the membrane CLI tool. It provides commands for logging in, searching for connectors, and executing actions against the Fidel API.
  • [PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it ingests and processes data from the external Fidel API which could potentially contain adversarial instructions.
  • Ingestion points: Data is received through CLI commands such as membrane action run, membrane action list, and membrane request as described in SKILL.md.
  • Boundary markers: There are no explicit delimiters or instructions provided to the agent to disregard content within the API responses that might resemble commands.
  • Capability inventory: The agent is equipped to execute shell commands (specifically the membrane CLI) and perform file operations.
  • Sanitization: The instructions do not specify any sanitization, filtering, or validation of the data retrieved from the Fidel service before the agent acts upon it.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 21, 2026, 11:47 PM