teleport-session-review

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

• Third-party content exposure detected (high risk: 0.85). Outsider free text from Teleport cluster data (e.g., resource_labels / resource_name / username fields in tctl recordings search --format=json) is parsed and inserted into the agent’s LLM context for the findings table and commentary, creating an indirect prompt-injection surface.