adobe-core-workflow-b
Pass
Audited by Gen Agent Trust Hub on Sep 12, 2026
Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
- [SAFE]: The skill instructions prioritize security best practices by requiring the redaction of bearer tokens (signed URLs), explicitly prohibiting the logging of sensitive input/output content, and mandating the prompt deletion of Adobe-hosted assets after processing.
- [INDIRECT_PROMPT_INJECTION]: The skill handles external PDF documents, creating a potential surface for indirect prompt injection. The workflow addresses this through a robust validation process. Ingestion points: Classifying and processing input PDF files described in SKILL.md. Boundary markers: Mandatory validation of media type, size, protection state, and operation contract before upload. Capability inventory: The skill uses Read, Glob, Grep, Write, and Edit tools to manage files and documentation within the local environment. Sanitization: Clear requirements to redact credentials and signed URLs from logs and to verify the structure and hashes of output artifacts.
Audit Metadata