moai-core-clone-pattern

Warn

Audited by Gen Agent Trust Hub on Mar 2, 2026

Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill requests and utilizes the Bash tool to perform complex, multi-step tasks across the codebase.\n
  • Evidence: SKILL.md lists Bash in allowed-tools.\n
  • Evidence: examples.md describes clones executing migrations and refactoring involving file modifications.\n- [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it ingests untrusted user data and passes it to autonomous clones with high-privilege tools.\n
  • Ingestion points: examples.md shows processing UserRequest objects in autonomous clones.\n
  • Boundary markers: No delimiters or protective instructions are used to separate user data from task logic.\n
  • Capability inventory: Includes Bash, Read, and Task tools, the latter of which allows for recursive agent spawning and independent execution with inherited permissions.\n
  • Sanitization: No sanitization or validation of the input payload is described before task delegation.
Audit Metadata
Risk Level
MEDIUM
Analyzed
Mar 2, 2026, 05:14 PM