deep-learning-experiment-workflow-skill

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

• Third-party content exposure detected (high risk: 0.90). The skill's Stage 1 Investigation explicitly requires reading and recording external sources (e.g., "reading open-source implementations", "checking papers", and the investigation-notes-template's "External URL"/"URLs" fields in stages/01-investigation/investigation-guide.md and investigation-notes-template.md), and those findings are required inputs driving downstream decisions and stage gates, so untrusted third-party content can materially influence agent actions.