public-relations

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

• Third-party content exposure detected (medium risk: 0.30). Runtime path ingests outsider-authored free text by reading inbound journalist request content from Connectively/Qwoted/Featured (daily digest/feed) in the “Respond to this HARO query” → “Daily Triage Workflow → Step 2/Read the request” steps.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill contains explicit runtime curl commands that fetch external content to feed the agent's detection loop (e.g., https://news.google.com/rss/search?q=QUERY&hl=en-US&gl=US&ceid=US:en), meaning remote data is pulled at runtime and would directly influence the agent's prompts/context.