Spec Driven Development
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Descripción del creador (original en inglés)
Security-hardened NotebookLM MCP with post-quantum encryption
Estos tres nunca se combinan en una sola puntuación: las opiniones, las mediciones y la popularidad son cosas distintas.
Todavía no escribimos un resumen sencillo para este. Abajo están los datos que leímos del registro.
Elige la app de IA que usas y copia la configuración.
Necesitas
¿Qué app usas?
Elige tu app arriba. La recordamos, así que la próxima vez el botón dirá “Instalar en …”.
Abre Claude Desktop → Settings → Developer → Edit Config y pega esto en claude_desktop_config.json
{
"mcpServers": {
"notebooklm-secure": {
"command": "npx",
"args": [
"-y",
"@pan-sec/notebooklm-mcp"
]
}
}
}Ejecuta esto en tu terminal
claude mcp add notebooklm-secure -- npx -y @pan-sec/notebooklm-mcpPégalo en .cursor/mcp.json de tu proyecto, o en ~/.cursor/mcp.json
{
"mcpServers": {
"notebooklm-secure": {
"command": "npx",
"args": [
"-y",
"@pan-sec/notebooklm-mcp"
]
}
}
}Tu navegador te pide permiso para abrir la app; luego la app muestra la configuración y te pide confirmar. Los valores de las claves quedan vacíos para que los completes tú.
Pégalo en .vscode/mcp.json de tu proyecto
{
"servers": {
"notebooklm-secure": {
"type": "stdio",
"command": "npx",
"args": [
"-y",
"@pan-sec/notebooklm-mcp"
]
}
}
}Tu navegador te pide permiso para abrir la app; luego la app muestra la configuración y te pide confirmar. Los valores de las claves quedan vacíos para que los completes tú.
Este se ejecuta en tu computadora, y ChatGPT solo puede conectarse a servidores en internet. Elige otra app.
Tu elección se guarda en una cookie en este dispositivo; no necesitas cuenta.
Esta configuración se genera a partir de la ficha del creador en el registro. Si el README dice otra cosa, sigue el README.
Se ejecuta en tu computadora; la lista de herramientas aparece en tu app de IA después de instalarlo.
Revisamos información que cualquiera puede ver; no hemos leído el código. Las revisiones que no aplican no se cuentan. · Revisado el 7 nov 2026
El nombre en el registro se verificó con la cuenta de GitHub Pantheon-Security, dueña del repo
Todavía no hemos leído este repo en GitHub
Todavía no hemos leído este repo en GitHub
Este se ejecuta en tu computadora; no hay endpoint en internet
Se ejecuta en tu computadora y usa ajustes locales en su lugar
Se ejecuta en tu computadora; solo podemos leer listas de herramientas de endpoints en internet
Comparado con todos los servidores activos de esta categoría en el registro. Las estrellas solo cuentan cuando quien lo publica es dueño del repo.
Superó 1 de 1 revisiones básicas aplicables. No hemos leído el código. Antes de agregar una clave o un token, revisa qué acceso pide el servidor y usa una clave que tenga solo los permisos que necesitas.
El registro no indica ninguna clave obligatoria, aunque algunos servidores te piden iniciar sesión la primera vez que los usas.
Se ejecuta en tu computadora. Tu app de IA lo inicia cuando hace falta, y los datos no pasan por los servidores del creador (a menos que el propio servidor llame a un servicio en línea).
El repo de GitHub cambió por última vez –, y la última versión (2026.5.0) llegó al registro hace 2 meses.
Las apps compatibles con MCP, como Claude Desktop, Claude Code, Cursor y VS Code. También necesitas Node.js en tu computadora.
En la misma categoría: Spec Driven Development, Remote Desktop Commander, Aiterm MCP
Servidores de la misma categoría hechos por otros creadores, ordenados por estrellas.
Cómo leer una tarjeta:★ Estrellas de GitHubiTres pilares: reseñas · revisiones · usoiActualizado = fecha en que se publicó la última versión en el registroi
MCP Server that facilitates spec-driven development workflows, not just Vibe Coding.
Remote Desktop Commander: use files, terminal and processes on your own computers from any AI app.
Persistent terminals and one launcher for Claude, Codex, Grok, Cursor, and Antigravity harnesses.
Copiado tal cual del repo del creador. PopNx MCP no lo ha revisado ni lo respalda. Se quitaron HTML, imágenes y scripts. Los enlaces llevan fuera de este sitio.
Zero-hallucination answers • Gemini Deep Research • 17 Security Layers • Enterprise Compliance
What's New 2026 • Deep Research • Document API • Create Notebooks • Security • Install
The only NotebookLM MCP with enterprise-grade security, post-quantum encryption, and full Gemini API integration.
Security-hardened fork of PleasePrompto/notebooklm-mcp • Maintained by Pantheon Security
Latest: v2026.3.1 — All 334 audit issues resolved. 631 tests. Full MCP protocol compliance.
| Version | Highlights |
|---|---|
| v2026.3.1 | Security Audit Complete — All 334 issues from the independent audit resolved. Tests: 609 → 631. Code quality: URL resolution deduplicated, handler extraction, non-null assertions eliminated. Test gaps closed: validateNotebookId, error body shape, delete_document confirm guard, sanitized throws, log rotation, rate-limiter memory bound, range clamping. |
| v2026.3.0 | The Security Audit Release — Four parallel AI code reviews (security, protocol, architecture, testing) against 334 issues. All highs and mediums resolved. Tests: 139 → 609 across 50 files (4.4×). Full MCP protocol compliance: structuredContent, isError, transport tags. Schema bounds on all 48 tools. Annotation correctness. Webhook SSRF fix. Audit log integrity (hash chain, concurrent write lock, rotation continuity). Per-page mutex. HandlerContext DI. Cert pinning retracted (claims aligned with implementation). |
| v2026.2.10 | The Hardening Release — 3 new security layers (14→17): secure-by-default auth, exponential backoff lockout, credential isolation. Architecture overhaul: 3,611-line handler split into 9 domain modules, tool registry pattern. Gemini API retry with backoff. Multi-stage Docker build. Token CLI (token show/rotate). 168 tests. |
| v2026.2.9 | performSetup no longer wipes credentials before Chrome opens — prevents auth destruction on failed launch |
| v2026.2.8 | cleanup_data never deletes auth dirs (browser_state/, chrome_profile/) — auth survives all cleanup paths |
| v2026.2.7 | Block headless setup_auth calls; robust auth-now.mjs standalone script with Chrome profile lock handling |
| v2026.2.6 | New add_folder tool for bulk PDF/file uploads; improved tier detection for NotebookLM Plus |
| v2026.2.5 | Fix show_browser silently ignored in setup_auth — browser now reliably opens for auth |
| v2026.2.4 | Auth stability: 7-day state expiry, touchStateFile on validation, block headless re_auth |
| v2026.2.3 | Studio Panel tools restored: Video Overview, Data Table generation re-enabled |
| v2026.2.1 | Standard profile expanded: 14 → 33 tools — notebook creation, audio, source management now visible by default |
| v2026.2.0 | Gemini 3 models (2.5 retiring March 31), Video Overviews, Data Tables, thinking level, structured JSON, SDK 1.41 |
| v2026.1.12 | Security hardening: constant-time auth, command injection fix, memory leak fixes, MCP SDK 1.26.0 |
| v2026.1.11 | Notebook sync extraction for new Angular UI, NOTEBOOKLM_NO_GEMINI env var |
| v2026.1.10 | Tool description clarity for multi-LLM compatibility |
| v2026.1.9 | Documentation & UX improvements, npm audit fix |
| v2026.1.8 | Major dependency updates (zod 4.x, dotenv 17.x, post-quantum 0.5.4) |
| v2026.1.7 | MCP Protocol UX: tool icons, human-friendly titles, behavior annotations |
| v2026.1.4 | Defense-in-depth path validation, security hardening |
| v2026.1.1 | Deep health checks, chat history extraction, context management |
# Quick install
claude mcp add notebooklm -- npx @pan-sec/notebooklm-mcp@latest
| Capability | Other MCPs | This MCP |
|---|---|---|
| Query NotebookLM | ✅ Basic | ✅ + session management, quotas |
| Create notebooks programmatically | ❌ | ✅ UNIQUE |
| Gemini Deep Research | ❌ | ✅ EXCLUSIVE |
| Document API (no browser) | ❌ | ✅ EXCLUSIVE |
| Post-quantum encryption | ❌ | ✅ Hybrid PQ at-rest |
| Enterprise compliance | ❌ | ✅ GDPR/SOC2/CSSF-ready |
| Video Overview generation | ❌ | ✅ NEW |
| Data Table extraction | ❌ | ✅ NEW |
| Chat history extraction | ❌ | ✅ |
| Deep health verification | ❌ | ✅ |
In April 2026, we commissioned a parallel deep-audit of v2026.2.11 (main @ 2973097) using four specialised AI code reviewers, each focused on a different attack surface: security vulnerabilities, protocol correctness, architecture quality, and testing gaps and edge cases. The four reviewers operated independently so their findings wouldn't influence each other. Together they produced a 334-item master issue list covering protocol correctness, security vulnerabilities, architecture flaws, test gaps, and documentation accuracy. All 334 issues are resolved across v2026.3.0 and v2026.3.1.
| Metric | Before (v2026.2.11) | After (v2026.3.1) |
|---|---|---|
| Tests | 139 | 631 across 50 files |
| Test suites | ~6 | 50 |
TypeScript errors (tsc --noEmit) | 0 | 0 (maintained) |
| npm audit vulnerabilities | 0 | 0 (maintained) |
| MCP protocol compliance | Partial | Full (structuredContent, isError, transport tags) |
| Audit log integrity | Basic | Hash-chain verified on read |
| Concurrent write safety | ❌ | ✅ Write-locked |
| Webhook SSRF | ❌ | ✅ Blocked |
Security reviewer: Identified the forceAuth bypass in validateToken() allowing unauthenticated access to filesystem tools; webhook SSRF via unvalidated delivery targets; audit log hash chain not verified on read; concurrent audit writes interleaving entries; auth token salt not persisted (tokens invalidated on restart).
Protocol reviewer: Found 38 tools returning incorrect response shapes (missing structuredContent, wrong isError semantics, transport tags leaking into content); all 48 tools had incorrect or missing readOnlyHint/idempotentHint/destructiveHint annotations; 9 tool schemas lacked numeric bounds, enabling out-of-range inputs.
Architecture reviewer: Flagged the 3,611-line handlers.ts as a maintenance liability; singleton imports throughout domain functions preventing unit testing; the 500-line switch/case dispatch adding O(n) overhead and making tool registration error-prone.
Testing & edge-case reviewer: Found test suite at 139 tests with minimal coverage of security-critical modules; mcp-auth.ts at near-zero coverage; no tests for prompt injection patterns, audit log tampering, or concurrent browser session state; the DSAR handler had an undetected race condition.
structuredContent/isError shapes; annotations accurate; schema bounds enforcedhandlers.ts split into 9 domain modules with HandlerContext dependency injection; 100% unit-testable without process mocks📋 Full Feature List (48 Tools)
| Tool | Description |
|---|---|
ask_question | Query notebooks with source-grounded answers |
add_notebook | Add a notebook to your library |
list_notebooks | List all notebooks in library |
select_notebook | Set active notebook |
update_notebook | Update notebook metadata |
remove_notebook | Remove from library |
create_notebook | Programmatically create new notebooks |
batch_create_notebooks | Create multiple notebooks at once |
sync_library | Sync library with NotebookLM |
list_sources | List sources in a notebook |
add_source | Add source to notebook |
add_folder | Bulk upload PDFs/files from a local folder to a notebook |
remove_source | Remove source from notebook |
generate_audio_overview | Create podcast-style audio |
get_audio_status | Check audio generation status |
download_audio | Download generated audio |
generate_video_overview | Create AI video overview (10 styles) |
get_video_status | Check video generation status |
generate_data_table | Generate structured data table |
get_data_table | Extract data table as JSON |
list_sessions | List active sessions |
close_session | Close a session |
reset_session | Reset session history |
get_health | Check server & auth status |
setup_auth | Initial authentication |
re_auth | Re-authenticate |
cleanup_data | Clean up local data |
get_quota | Check usage quotas |
set_quota_tier | Set quota tier |
get_query_history | View past queries |
get_notebook_chat_history | Extract browser chat history |
| get_project_info | Get projec
Solo se muestra el principio; lee el resto en el repo