why

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

• Third-party content exposure detected (high risk: 0.85). The skill’s runtime workflow explicitly queries multiple MCP-backed sources (issue tracker, long-form docs, Slack, Datadog, Sentry, Databricks) and then synthesizes their returned text into the agent’s LLM context, so outsider-authored free text from those systems can be ingested via the MCP results.