themodernsoftware-notebooklm

MEDIUM W011: Third-party content exposure detected (indirect prompt injection risk).

• Third-party content exposure detected (high risk: 1.00). The skill's required workflow (SKILL.md Step 2 and related crawl rules) instructs the agent to crawl and ingest public pages from https://themodernsoftware.dev/ using agent-browser and add those discovered URLs/files into NotebookLM, so untrusted third‑party page content will be read and can directly influence generation and downstream actions.

MEDIUM W012: Unverifiable external dependency detected (runtime URL that controls agent).

• Potentially malicious external URL detected (high risk: 0.80). The skill explicitly opens and crawls https://themodernsoftware.dev/ at runtime and ingests fetched pages into NotebookLM as sources that are injected into the model context to drive generation, making that external site a required runtime dependency that directly controls agent outputs.