The Agent Skills Directory

[COMMAND_EXECUTION]: The skill invokes local Python scripts (scripts/extract_transcript.py, scripts/extract_article.py, scripts/transcribe_audio.py, scripts/generate_html.py, and scripts/fetch_images.py) to handle specialized tasks like content extraction and output formatting.- [EXTERNAL_DOWNLOADS]: The engine utilizes the WebSearch tool to locate and verify stock images from well-known services including Pixabay, Unsplash, and Pexels to support visual content generation.- [PROMPT_INJECTION]: The skill processes untrusted data from external URLs and local files as part of its core atomization and orchestration workflow. This creates a surface for indirect prompt injection as defined in Category 8.
Ingestion points: Content is ingested through URL extraction scripts and the Read tool as specified in SKILL.md.
Boundary markers: The instructions do not explicitly include delimiters or 'ignore' directives for the source content when passed to sub-agents.
Capability inventory: The skill possesses the ability to execute subprocesses, write files to the local filesystem, and perform network searches.
Sanitization: There are no documented sanitization or filtering steps for the ingested content before it is processed by the orchestration agents.

repurpose