themodernsoftware-notebooklm

Pass

Audited by Gen Agent Trust Hub on Feb 25, 2026

Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [EXTERNAL_DOWNLOADS]: The skill crawls course materials from https://themodernsoftware.dev/ as its primary function to generate study outputs.\n- [COMMAND_EXECUTION]: Local shell scripts scripts/init-workspace.sh and scripts/verify-week-state.sh perform file system operations and data validation using jq. The skill also invokes agent-browser and notebooklm CLI tools to execute its defined workflow.\n- [PROMPT_INJECTION]: The skill processes external web content, creating a potential surface for indirect prompt injection. \n
  • Ingestion points: Content is crawled from the external domain themodernsoftware.dev using agent-browser.\n
  • Boundary markers: Prompt templates in templates/prompts/ lack explicit delimiters or safety instructions for handling the ingested text.\n
  • Capability inventory: Operations are limited to notebooklm ingestion and generation, and local workspace management scripts.\n
  • Sanitization: No explicit sanitization or validation of the crawled content is performed prior to processing.
Audit Metadata
Risk Level
SAFE
Analyzed
Feb 25, 2026, 05:44 AM