scholar-agent

Pass

Audited by Gen Agent Trust Hub on Apr 17, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses the subprocess module to interact with external CLI tools and bundled shell scripts.
  • It executes playwright-cli (an Anthropic tool) for session management and browser automation tasks related to Scholar Inbox and NotebookLM.
  • It utilizes system utilities like pgrep, ps, and pkill to manage browser profiles and handle lock conflicts.
  • It executes bundled shell scripts (e.g., add_to_notebooklm.sh) located within the skill's directory for notebook management tasks.
  • [EXTERNAL_DOWNLOADS]: The skill facilitates interaction with external services and recommends the installation of third-party tools.
  • It communicates with the Scholar Inbox API (api.scholar-inbox.com) using the Python standard library's urllib.request to fetch paper data and submit user feedback.
  • It relies on notebooklm-py and playwright-cli for advanced functionality, both of which are standard tools in their respective ecosystems.
  • [CREDENTIALS_UNSAFE]: The skill handles session cookies for authentication with Scholar Inbox.
  • It provides mechanisms to extract or manually input session cookies, which are then stored locally at ~/.config/scholar-inbox/session.json.
  • The implementation follows security best practices by setting file permissions to 0600 (read/write only by the owner) to protect the stored session data.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 17, 2026, 03:17 AM