still-image-retouching-finishing
Pass
Audited by Gen Agent Trust Hub on Aug 20, 2026
Risk Level: SAFE
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill references technical documentation and standards from established organizations including Adobe, the International Color Consortium (ICC), CIE, IPTC, and the C2PA. These URLs are provided as educational resources and grounded facts and do not involve automated downloads or execution of external assets.
- [PROMPT_INJECTION]: No patterns of prompt injection, role-play bypasses, or instructions to ignore safety guidelines were detected. The skill includes specific constraints to prevent its use for inappropriate tasks like generating synthetic images from scratch or performing unethical alterations in documentary contexts.
- [DATA_EXFILTRATION]: There are no hardcoded credentials, API keys, or instructions that suggest the collection or unauthorized transmission of sensitive user data. The instructions specifically advise on stripping sensitive metadata like GPS and serial numbers when necessary for privacy.
- [REMOTE_CODE_EXECUTION]: The skill is entirely descriptive and tool-agnostic. It does not contain any scripts, package installation commands (npm, pip), or shell execution patterns (curl | bash).
- [COMMAND_EXECUTION]: No usage of the dynamic context injection syntax (
!command) or arbitrary shell commands was found in the documentation.
Audit Metadata