adobe-data-handling
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill establishes a governance process for managing Adobe data workflows. It consists of instructional guidance without providing executable scripts or dangerous shell commands.
- [DATA_EXFILTRATION]: The skill instructions mandate treating authentication tokens and signed URLs as secrets. It explicitly prohibits exposing these sensitive identifiers in logs, tickets, or analytics, adhering to security best practices.
- [EXTERNAL_DOWNLOADS]: The skill references official Adobe developer documentation for Firefly, PDF Services, and Privacy Service APIs. These references point to well-known service domains and do not involve untrusted remote code execution.
- [INDIRECT_PROMPT_INJECTION]: The skill identifies a potential attack surface through the processing of Adobe workflow data.
- Ingestion points: The skill processes Adobe workflow inputs, including PDFs, extracted text, and Firefly prompts.
- Boundary markers: The instructions require mapping all storage hops, jurisdictions, and recipients to maintain custody boundaries.
- Capability inventory: Tools are limited to Read, Glob, Grep, Write, and Edit for repository artifact inspection.
- Sanitization: The workflow requires testing redaction, access expiry, and content-policy outcomes to ensure data is handled correctly before any live side effects.
Audit Metadata