internal-comms

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

• Third-party content exposure detected (medium risk: 0.30). In examples/3p-updates.md (and similarly examples/company-newsletter.md / examples/faq-answers.md), the runtime workflow’s “Tools Available”/“Tools to use” directs the agent to pull from Slack and Email content, which is outsider-authored free text that could be posted to channels/inboxes the agent reads.