The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses subprocess.run in scripts/apply_cuts.py to interact with the local screenkite-alpha CLI tool. This is performed using a list of arguments rather than a shell string, which is a secure practice that prevents command injection.
[EXTERNAL_DOWNLOADS]: The skill uses uv to manage the requests library dependency. requests is a standard, well-known package used for the transcription tasks. This is considered safe as it utilizes a trusted package registry and established tooling.
[CREDENTIALS_UNSAFE]: The documentation mentions using ELEVEN_LABS_API_KEY but correctly instructs the user or agent to manage it via environment variables or .env files rather than hardcoding. This aligns with security best practices for secret management.
[SAFE]: The core logic in scripts/compute_all_cuts.py involves processing JSON-formatted transcription data to identify cut points. The script uses safe string processing and JSON parsing, with no evidence of dynamic code execution (like eval or exec).

screenkite-clean-cut