there-there

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

• Third-party content exposure detected (high risk: 0.80). The skill's workflows and command reference (SKILL.md and references/commands.md) show the CLI fetching ticket messages via there-there API (there-there show-ticket returns messages[] with body_html and inline_images that come from customers), and the triage/respond workflows explicitly instruct the agent to read those user-generated messages and then take actions (reply, assign, close), so untrusted third-party content can materially influence behavior.