human-social

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

• Third-party content exposure detected (low risk: 0.10). The workflow fetches guidelines from a designated URL via WebFetch, which constitutes reading an external web resource rather than consuming outsider-submitted messages directly.

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

• Potentially malicious external URL detected (high risk: 0.70). The skill instructs fetching behavior-selecting guidelines from an arbitrary, unverified domain (guidelines.sh), constituting a weak operational dependency.