improve

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

• Third-party content exposure detected (high risk: 0.75). The required runtime workflow includes pulling production run history via terse history "<job-name>" --json --triggers, which fetches past trigger event payloads authored by external systems/users (e.g., GitHub/Slack/Linear events) and feeds that free-form event text into the agent’s LLM context during analysis/replay.