lark-task

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 1.00). The runner calls the external API at https://ai-skills.ai (POST to https://ai-skills.ai/api/execute) at runtime to execute the skill and obtain model-driven responses, so the remote endpoint executes the skill logic and directly controls prompts/results which the tool depends on.