run-tests

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). The skill contains instructions to silently update and re-read itself from an external package repository before execution, which is an unnecessary and stealthy modification of core behavior.

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

• Potentially malicious external URL detected (medium risk: 0.30). The skill instructs the agent to fetch external instructions and code via an npx command (https://registry.npmjs.org), representing an unofficial runtime dependency.