The Agent Skills Directory

[COMMAND_EXECUTION]: The skill includes a Python script scripts/validate_seedance_segment_map.py used to validate the structure of editing plans. Analysis of the source code confirms it is a benign utility that uses standard libraries (json, argparse, pathlib) to perform data validation without any side effects like network access or file modifications.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and analyze untrusted data such as script drafts and prompt configurations. However, it implements strict boundary markers through the '七段式' (Seven-part) structure and uses automated validation scripts to enforce schema compliance, which significantly mitigates the risk of the agent following instructions embedded in the analyzed data.
[DATA_EXFILTRATION]: No evidence of network communication or unauthorized access to sensitive local files (e.g., credentials, SSH keys) was found in any of the provided instructions or scripts.

sd-qc-engine