The Agent Skills Directory

[COMMAND_EXECUTION]: The skill utilizes local Python scripts, specifically scripts/seedance_lint.py and scripts/higgsfield_memory.py, to perform pre-flight checks on prompts and to log performance data. These executions are confined to the repository's local environment and serve the skill's diagnostic functions.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-provided scene descriptions. To mitigate risks associated with untrusted input, it mandates the use of a linter script that screens for policy violations, restricted brand names, and sensitive content before a prompt is finalized for generation.
[SAFE]: The documentation is thorough, professionally written, and adheres to cinematic standards. All external references point to well-known technology domains associated with the platform's infrastructure (e.g., BytePlus, ByteDance). No obfuscation or credential harvesting was identified.

higgsfield-seedance