The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill installs the notebooklm-py package and the Playwright Chromium browser using the uv tool. The notebooklm-py tool is a resource associated with the skill author. Evidence: uv tool install --with playwright --force notebooklm-py and $(uv tool dir)/notebooklm-py/bin/playwright install chromium in SKILL.md.
[COMMAND_EXECUTION]: The skill relies on various command-line tools to perform its functions, including uv, notebooklm, playwright, and jq. Evidence: Multiple shell blocks in SKILL.md demonstrating the use of notebooklm create, notebooklm source add, and notebooklm generate.
[INDIRECT_PROMPT_INJECTION]: The skill ingests data from external source documents to generate slide content and titles, creating a surface for indirect instructions. Ingestion points: External document provided at <doc-path> (referenced in SKILL.md). Boundary markers: None. Capability inventory: Shell command execution (notebooklm), network communication, and document generation. Sanitization: Document metadata is wrapped in double quotes when passed to shell commands, but no further sanitization of shell-active characters (like backticks) is specified.
[DYNAMIC_EXECUTION]: The skill constructs and executes shell commands at runtime by interpolating strings extracted from the source document (e.g., H1 headers or filenames). Evidence: NOTEBOOK_ID=$(notebooklm create "<title>" --json | jq -r .notebook.id) in SKILL.md, where <title> is derived from the source document.

wowerpoint