axiom-validate-screenshots

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

• Third-party content exposure detected (medium risk: 0.30). The required runtime workflow reads user-provided image files from a user-specified folder path (Step 2 glob + Step 4 Read tool), and those images can contain outsider-authored free text (e.g., placeholder/debug/competitor strings) that the LLM ingests for analysis.