The Agent Skills Directory

[COMMAND_EXECUTION]: The skill provides functionality via the lj ingest command and lj ingest codex tools to execute shell commands and capture their output or terminal recordings. These are used to generate 'evidence' artifacts for the project's auditable production trail.
Evidence found in scripts/providers/terminal_record_cli.py where user-supplied command strings are executed via subprocess.run after being parsed by shlex.split.
Evidence in packages/core/credentials.py where OS-native tools like security (macOS) and secret-tool (Linux) are used to securely manage API credentials.
[EXTERNAL_DOWNLOADS]: The skill facilitates connections to external cloud providers for AI services (TTS and video generation).
Evidence in scripts/providers/seedance_scene_cli.py and providers/volcengine_tts.py which perform network requests to Volcengine (Bytedance) API endpoints.
Evidence in providers/openai_compatible.py which connects to user-configured OpenAI-compatible endpoints.
The skill utilizes npx for dynamic execution of the hyperframes utility as documented in packages/core/visual_generation.py.

lingjian-video