The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill instructs the user to download a docker-compose.yml configuration file from the project's public GitHub repository (https://raw.githubusercontent.com/lfnovo/open-notebook/main/docker-compose.yml) for deployment. This is a standard and transparent installation method for the tool.
[PROMPT_INJECTION]: The skill implements a research knowledge base that processes untrusted external data, creating a surface for Indirect Prompt Injection.
Ingestion points: Untrusted data enters the system through the /api/sources endpoint, supporting web URLs, PDF uploads, and raw text (referenced in SKILL.md and references/api_reference.md).
Boundary markers: The documentation and example prompts do not specify the use of delimiters or 'ignore embedded instructions' warnings for processed context.
Capability inventory: The system has capabilities to write to a database (notes, credentials), generate audio files (podcasts), and perform network operations via AI providers.
Sanitization: There is no evidence of content sanitization or instruction filtering in the provided integration scripts.
[COMMAND_EXECUTION]: The test suite (scripts/test_open_notebook_skill.py) utilizes the Python compile() function to verify that example scripts are syntactically correct. This is a controlled use of dynamic execution for testing purposes and does not execute untrusted input from the network.

open-notebook