competitive-analysis

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

• Third-party content exposure detected (high risk: 0.80). The skill's required workflow (SKILL.md Step 3 "collect evidence", the Monitoring Plan, and references like G2/Reddit/Hacker News/Crunchbase) explicitly instructs the agent to gather and consume publicly available pages and user-generated content (pricing pages, reviews, forum posts), which the agent is expected to read and use to drive analysis and actions—exposing it to untrusted third-party content that could carry indirect prompt injections.