The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches the official Docker Compose configuration from GitHub to deploy the self-hosted environment. This is a standard deployment procedure for the application and uses a well-known service for the download.
[COMMAND_EXECUTION]: The skill requires the use of docker compose and Bash to manage the local application stack. These commands are necessary for the self-hosted infrastructure and are restricted to the local execution environment.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted research materials such as PDFs, Office documents, and web pages to provide context for AI-generated summaries and chat interactions.
Ingestion points: External data enters the system context through the /api/sources endpoint, as documented in SKILL.md and scripts/source_ingestion.py.
Boundary markers: The instructions do not define explicit delimiters or instructions to prevent the model from following commands embedded within retrieved documents.
Capability inventory: The skill can execute local Python scripts that interact with a REST API to perform content management and chat execution.
Sanitization: Content processing is managed by the underlying FastAPI and LangChain services; the skill documentation does not provide specific client-side sanitization rules.
[DYNAMIC_EXECUTION]: The test suite in scripts/test_open_notebook_skill.py uses the Python compile() function. This is used exclusively for syntax validation of code examples during testing and does not execute untrusted code at runtime.

alterlab-open-notebook