The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses a bundled Python script, scripts/lint_prompt.py, to analyze the structure of generated prompts. Technical analysis of the script confirms it is a self-contained text-processing utility that uses standard libraries for regex-based validation. It does not perform network operations or execute dynamic code.
[EXTERNAL_DOWNLOADS]: Documentation references official Lark documentation for Seedance 2.5 and an informational GitHub repository. These references are used neutrally to verify model capabilities and design constraints without initiating untrusted remote execution.
[SAFE]: The skill manages the processing of user-supplied data through a dedicated pipeline that includes a structural extraction phase and a validation step. 1. Ingestion points: User production ideas and multimodal asset descriptions processed in SKILL.md. 2. Boundary markers: Explicit task templates and role-mapping requirements in task-patterns.md. 3. Capability inventory: Execution of the local Python linter. 4. Sanitization: Heuristic structural verification performed by the lint_prompt.py script.

seedance-25