The Agent Skills Directory

[DYNAMIC_EXECUTION]: The test script scripts/test_open_notebook_skill.py uses the Python compile() function.\n
Evidence: compile(content, filename, "exec") in the _check_script method.\n
Context: This is a benign implementation used to verify the syntax of the provided example Python scripts during development or testing.\n- [REMOTE_CODE_EXECUTION]: The skill documentation includes instructions to download an external configuration file via curl.\n
Evidence: curl -o docker-compose.yml https://raw.githubusercontent.com/lfnovo/open-notebook/main/docker-compose.yml in SKILL.md.\n
Context: The download points to the official repository of the Open Notebook project and represents a standard deployment procedure.\n- [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process external content such as web pages and PDF documents.\n
Ingestion points: API endpoints for creating sources (/api/sources) defined in SKILL.md and references/api_reference.md.\n
Boundary markers: None explicitly defined in the skill's API client code.\n
Capability inventory: Uses the requests library to communicate with the self-hosted Open Notebook API for document analysis and chat functions.\n
Sanitization: Input validation and sanitization are managed by the self-hosted backend application.\n- [DATA_EXPOSURE_AND_EXFILTRATION]: The skill requires API credentials for various AI providers to function.\n
Evidence: Usage of api_key in credential creation API examples.\n
Context: The documentation correctly utilizes placeholders (e.g., sk-...) and recommends using environment variables for sensitive configuration, following security best practices.

open-notebook