forge-harmonize
Pass
Audited by Gen Agent Trust Hub on Apr 16, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious patterns detected. The skill logic is confined to deterministic color calculations based on provided brand colors.\n- [DATA_EXFILTRATION]: The skill does not access sensitive files or perform network operations.\n- [REMOTE_CODE_EXECUTION]: No external dependencies, remote scripts, or dynamic code generation are involved.\n- [COMMAND_EXECUTION]: The skill requests writing a generated palette to a local configuration file (.forge.json), which is consistent with its stated purpose of managing project design assets.\n- [PROMPT_INJECTION]: The instructions are task-oriented and do not attempt to bypass agent constraints or safety filters.
Audit Metadata