notebooklm-slides
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFE
Full Analysis
- COMMAND_EXECUTION (SAFE): The skill invokes a local script (start-server.sh) from a sibling skill directory and uses standard shell commands (mv) to organize files. These operations are transparent, use defined paths, and are integral to the stated workflow.
- EXTERNAL_DOWNLOADS (SAFE): The skill's primary function involves interacting with notebooklm.google.com, which is a legitimate and expected service for the described task. No unauthorized or suspicious network connections were identified.
- INDIRECT_PROMPT_INJECTION (SAFE): The skill processes untrusted educational content (lessons, quizzes) for upload. Ingestion points: local educational files. Boundary markers: none. Capability inventory: browser interaction and shell file operations. Sanitization: none. Although markers are absent, the risk is negligible as the data is consumed by a third-party tool for document generation and not interpreted as instructions by the agent.
Audit Metadata