media-higgsfield-assets
Pass
Audited by Gen Agent Trust Hub on Aug 19, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill interfaces with external MCP tools via a defined protocol (
models_explore,get_cost,job_status). These are restricted to platform-approved actions for exploring model capabilities and managing job lifecycles. - [PROMPT_INJECTION]: The skill includes instructions to prevent misuse, such as prohibiting the cloning of voices without consent and warning against guessing model parameters. It enforces strict boundary rules by requiring the agent to use live schemas rather than hardcoded assumptions.
- [EXTERNAL_DOWNLOADS]: The skill references the official Higgsfield Skills GitHub repository (
github.com/higgsfield-ai/skills) for documentation and logic benchmarks. This is a recognized vendor resource for the service being integrated.
Audit Metadata