The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses python3 scripts/reel_text.py to calculate caption timings. The script is included in the skill package, uses only the Python standard library, and performs safe JSON processing without shell injection risks or external network calls.
[INDIRECT_PROMPT_INJECTION]: The skill processes user-provided transcripts and scene data. While these are ingestion points for external data, the reel_text.py script treats this data as strictly structured JSON for geometry and timing calculations. It does not interpolate this data back into LLM prompts in a way that could trigger instruction override, and it focuses on rendering text onto video layers.
[REMOTE_CODE_EXECUTION]: No remote code execution patterns were detected. All logic is contained within the provided local script.
[DATA_EXFILTRATION]: No network operations or sensitive file access patterns were found. The script operates entirely offline and only prints the modified scene JSON to standard output.

spoken-captions