learn

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). The skill's workflow (Step 0 and the "If URL (http* or owner/repo format)" and agent prompts) explicitly clones and reads arbitrary GitHub repositories via ghq and then instructs agents to "READ source code from: [SOURCE_DIR]"—i.e., ingesting untrusted, user-generated third-party content that the agents must interpret and that can influence subsequent actions.

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

• Potentially malicious external URL detected (high risk: 0.90). The skill runs ghq get -u on user-supplied GitHub URLs (e.g. https://github.com/$OWNER/$REPO) at runtime and then instructs agents to "READ source code from" the cloned SOURCE_DIR, which injects external repository content directly into agent prompts/contexts and is a required dependency for the skill.