The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the auto_label command-line interface to execute the video annotation and distillation workflow.
[REMOTE_CODE_EXECUTION]: The skill implements a prompts_module configuration that dynamically loads and executes Python modules from the system's PYTHONPATH at runtime to support domain-specific customization.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection as it ingests untrusted external data.
Ingestion points: The pipeline reads video files from video_root and structured data from input_jsonl_files (referenced in SKILL.md).
Boundary markers: No explicit delimiter wrappers or instructions to ignore embedded content were found in the prompt templates.
Capability inventory: The skill has the ability to execute shell commands via auto_label and perform network operations to communicate with VLM/LLM endpoints.
Sanitization: The skill does not sanitize or filter the content of processed videos or input JSONL files before passing them to model prompts.

tao-generate-video-reasoning-annotations