Skills
skills/modu-ai/moai-adk/moai-workflow-worktree/Gen Agent Trust Hub

moai-workflow-worktree

Pass

Audited by Gen Agent Trust Hub on Mar 10, 2026

Risk Level: SAFECOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONPROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses Python's subprocess module to execute Git and CLI commands for managing worktree states and registry metadata as seen in references/examples.md.
  • [REMOTE_CODE_EXECUTION]: The template system described in modules/moai-adk-integration.md supports setup_commands, which executes user-defined shell commands during the initialization of new worktree environments.
  • [PROMPT_INJECTION]: There is an indirect prompt injection surface (Category 8) where SPEC identifiers and descriptions are interpolated into shell command strings. Ingestion points: SPEC IDs and descriptions in SKILL.md and modules/worktree-commands.md. Boundary markers: Absent. Capability inventory: subprocess.run (Python) and shell eval (CLI). Sanitization: Absent; metadata is interpolated directly without documented shell escaping.
  • [EXTERNAL_DOWNLOADS]: Fetches and installs the moai tool from the vendor's GitHub repository (github.com/modu-ai/moai-adk) for CI/CD and setup purposes as documented in references/reference.md.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 10, 2026, 07:39 PM