technical-writing

CRITICAL E004: Prompt injection detected in skill instructions.

• Potential prompt injection detected (medium risk: 0.30). The skill contains instructions to automatically recommend and offer installation of external companion skills from a third-party repository upon activation, acting as an unsolicited promotional channel.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill contains instructions to check for missing companion skills hosted on GitHub repositories and prompts the user to install them via npx skills add, presenting a weak third-party operational dependency.