mark-evidence

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

• Third-party content exposure detected (high risk: 0.80). This skill explicitly reads and parses untrusted, user-generated content from PR titles/bodies and issue-level comments (Step 2 and Step 4a), scans Linear issue bodies/comments via Linear MCP (Step 4b) and extracts Slack/Notion links and identifier tokens (Steps 3, 4, 7) which are then recorded in the Manifest and drive downstream processing, so third-party content can materially influence subsequent tool use.