suppressing-noisy-errors

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

• Third-party content exposure detected (high risk: 0.70). The required workflow uses PostHog API tools to pull live error/issue records (e.g., issue description/sample messages and $exception stack traces) into the agent’s context, where those fields are outsider-authored event data (from third-party user traffic and extensions/network sources), creating an indirect prompt-injection pathway.