competitive-analysis

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

• Third-party content exposure detected (high risk: 0.95). Required workflow uses WebSearch/WebFetch to fetch App Store pages and competitor websites (public web content authored by outsiders), which is then ingested into the agent’s LLM context for analysis.

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

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly instructs using WebSearch/WebFetch at runtime to fetch competitor pages (e.g., https://todoist.com/features and https://todoist.com/pricing), and that fetched page content is injected into the agent's analysis flow to drive prompts and outputs, so these runtime fetches constitute an external dependency that directly controls the agent's behavior.