midscene-yaml-generator

Pass

Audited by Gen Agent Trust Hub on Mar 4, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the Bash tool to execute local Node.js scripts (scripts/health-check.js and scripts/midscene-run.js) to verify the environment and execute generated automation workflows.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it transforms untrusted user-provided natural language into executable YAML configurations. These configurations have the capability to execute shell commands (external_call: shell) and JavaScript (javascript:), which could be exploited if a user provides malicious instructions.
  • [PROMPT_INJECTION]: \n
  • Ingestion points: User-provided natural language descriptions of automation tasks. \n
  • Boundary markers: None identified; the skill processes the raw input string to generate structured YAML. \n
  • Capability inventory: Access to file system tools (Read, Write, Edit, Bash) and the ability to execute generated shell and JavaScript code through the automation runner. \n
  • Sanitization: Includes an explicit 'Safety Considerations' section that guides the AI to avoid generating destructive shell commands, prevent SSRF via external_call: http, and avoid exposing sensitive environment variables.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 4, 2026, 09:19 PM