vtcode-config 0.133.22

# ============================================================
# VT Code — Agent Network Allowlist
# network_allowlist.toml
#
# Curated safe domains for AI agent egress filtering.
# Suitable for macOS Seatbelt profiles, Linux Landlock/seccomp
# rules, or embedding in VT Code's agent config schema.
#
# Sections:
#   [ai_providers]       — LLM inference endpoints (cloud + local)
#   [search]             — Web & specialized search APIs
#   [web_crawl]          — Full-page fetch & content extraction
#   [mcp_servers]        — MCP tool server endpoints
#   [package_registries] — Language package managers
#   [code_hosting]       — Git platforms & raw content
#   [auth]               — OAuth & identity providers
#   [dev_infra]          — Cloud infra, databases, observability
#   [os_updates]         — System package mirrors (Ubuntu)
#
# Usage (Rust):
#   let config: NetworkAllowlist = toml::from_str(include_str!("network_allowlist.toml"))?;
# ============================================================

[meta]
version      = "1.1.0"
last_updated = "2026-06-27"
maintainer   = "vinhnx"
repo         = "github.com/vinhnx/VTCode"
description  = "Curated egress allowlist for VT Code agent sandboxing"

# ============================================================
# AI MODEL PROVIDERS — CLOUD
# ============================================================

[[ai_providers.cloud]]
name     = "Anthropic"
domain   = "api.anthropic.com"
protocol = "https"
notes    = "Claude models; primary Anthropic Messages API"

[[ai_providers.cloud]]
name     = "OpenAI"
domain   = "api.openai.com"
protocol = "https"
notes    = "GPT-4o, o3, o4-mini series"

[[ai_providers.cloud]]
name     = "Google GenAI"
domain   = "generativelanguage.googleapis.com"
protocol = "https"
notes    = "Gemini via Google AI Studio"

[[ai_providers.cloud]]
name     = "Google Vertex AI"
domain   = "us-central1-aiplatform.googleapis.com"
protocol = "https"
notes    = "Gemini + third-party models via Vertex"

[[ai_providers.cloud]]
name     = "Mistral"
domain   = "api.mistral.ai"
protocol = "https"
notes    = "Mistral Large, Codestral"

[[ai_providers.cloud]]
name     = "Groq"
domain   = "api.groq.com"
protocol = "https"
notes    = "Ultra-fast LPU inference; Llama, Mixtral"

[[ai_providers.cloud]]
name     = "Cohere"
domain   = "api.cohere.com"
protocol = "https"
notes    = "Command R+, embeddings, rerank"

[[ai_providers.cloud]]
name     = "Together AI"
domain   = "api.together.xyz"
protocol = "https"
notes    = "OpenAI-compatible multi-model hosting"

[[ai_providers.cloud]]
name     = "Fireworks AI"
domain   = "api.fireworks.ai"
protocol = "https"
notes    = "Fast open model serving; FireFunction"

[[ai_providers.cloud]]
name     = "DeepSeek"
domain   = "api.deepseek.com"
protocol = "https"
# ACTION REQUIRED: migrate model strings before late-Jul 2026
notes    = "DeepSeek V3, R1"

[[ai_providers.cloud]]
name     = "xAI (Grok)"
domain   = "api.x.ai"
protocol = "https"
notes    = "Grok-3 series"

[[ai_providers.cloud]]
name     = "Perplexity"
domain   = "api.perplexity.ai"
protocol = "https"
notes    = "Search-augmented models"

[[ai_providers.cloud]]
name     = "GitHub Copilot"
domain   = "api.githubcopilot.com"
protocol = "https"
notes    = "Official copilot-language-server SDK path; BYOK-compliant"

[[ai_providers.cloud]]
name     = "OpenRouter"
domain   = "openrouter.ai"
protocol = "https"
notes    = "Unified multi-provider routing"

[[ai_providers.cloud]]
name     = "Novita AI"
domain   = "api.novita.ai"
protocol = "https"
notes    = "Open model hosting; serverless inference"

[[ai_providers.cloud]]
name     = "Cerebras"
domain   = "api.cerebras.ai"
protocol = "https"
notes    = "Wafer-scale fast inference"

[[ai_providers.cloud]]
name     = "StepFun"
domain   = "api.stepfun.com"
protocol = "https"
notes    = "Step-3.7 Flash; applied to Startup Program"

[[ai_providers.cloud]]
name     = "Zhipu AI (GLM)"
domain   = "open.bigmodel.cn"
protocol = "https"
notes    = "GLM-5.2 series"

[[ai_providers.cloud]]
name     = "MiniMax"
domain   = "api.minimax.chat"
protocol = "https"
notes    = "MiniMax-Text-01"

[[ai_providers.cloud]]
name     = "Moonshot (Kimi)"
domain   = "api.moonshot.cn"
protocol = "https"
notes    = "Long-context models"

[[ai_providers.cloud]]
name     = "Baichuan"
domain   = "api.baichuan-ai.com"
protocol = "https"
notes    = "Baichuan-4 series"

[[ai_providers.cloud]]
name     = "01.AI (Yi)"
domain   = "api.lingyiwanwu.com"
protocol = "https"
notes    = "Yi-Large"

[[ai_providers.cloud]]
name     = "SambaNova"
domain   = "api.sambanova.ai"
protocol = "https"
notes    = "RDU fast inference"

[[ai_providers.cloud]]
name     = "Poolside"
domain   = "api.poolside.ai"
protocol = "https"
notes    = "Laguna model; code-focused"

[[ai_providers.cloud]]
name     = "MiMo"
domain   = "api.mimo.ai"
protocol = "https"
verify   = true
# ACTION REQUIRED: migrate MiMo V2.5 Pro/Flash model strings before end-Jun 2026
notes    = "MiMo V2.5 Pro/Flash"

[[ai_providers.cloud]]
name     = "Atlas Cloud"
domain   = "api.atlas.ai"
protocol = "https"
verify   = true
notes    = "Verify domain before use"

[[ai_providers.cloud]]
name     = "EvoLink AI"
domain   = "api.evokeai.com"
protocol = "https"
verify   = true
notes    = "Verify domain before use"

# ============================================================
# AI MODEL PROVIDERS — LOCAL / SELF-HOSTED
# ============================================================

[[ai_providers.local]]
name     = "LM Studio"
host     = "localhost"
port     = 1234
protocol = "http"
notes    = "OpenAI-compatible local server"

[[ai_providers.local]]
name     = "Ollama"
host     = "localhost"
port     = 11434
protocol = "http"
notes    = "Ollama REST API"

[[ai_providers.local]]
name     = "LlamaCpp"
host     = "localhost"
port     = 8080
protocol = "http"
notes    = "llama.cpp default HTTP server port"

# ============================================================
# SEARCH — WEB
# ============================================================

# Recommended primary for VT Code: agent-optimized, clean DX
[[search.web]]
name     = "Tavily"
domain   = "api.tavily.com"
protocol = "https"
notes    = "Agent-optimized; best developer UX for RAG + tool-use stacks"

# Neural/semantic search — good for code and research queries
[[search.web]]
name     = "Exa AI"
domain   = "api.exa.ai"
protocol = "https"
notes    = "Neural semantic search"

# Privacy-respecting — aligns with VT Code's local-first ethos
[[search.web]]
name     = "Brave Search"
domain   = "api.search.brave.com"
protocol = "https"
notes    = "Privacy-first; no user tracking"

[[search.web]]
name     = "Serper"
domain   = "google.serper.dev"
protocol = "https"
notes    = "Google SERP results; cost-effective"

[[search.web]]
name     = "SerpAPI"
domain   = "serpapi.com"
protocol = "https"
notes    = "Multi-engine: Google, Bing, DuckDuckGo"

[[search.web]]
name     = "ValueSERP"
domain   = "api.valueserp.com"
protocol = "https"
notes    = "Google SERP scraper API"

[[search.web]]
name     = "DataForSEO"
domain   = "api.dataforseo.com"
protocol = "https"
notes    = "SERP data + keyword intelligence"

[[search.web]]
name     = "Bing Search"
domain   = "api.bing.microsoft.com"
protocol = "https"
notes    = "Microsoft Azure Cognitive Search"

[[search.web]]
name     = "You.com"
domain   = "api.you.com"
protocol = "https"
notes    = "LLM-native search results"

[[search.web]]
name     = "Kagi"
domain   = "kagi.com"
protocol = "https"
notes    = "High signal-to-noise, ad-free"

# No full web search API — Instant Answers only; for full SERP use via SerpAPI wrapper
[[search.web]]
name     = "DuckDuckGo"
domain   = "api.duckduckgo.com"
protocol = "https"
notes    = "Instant Answers API; privacy-first, no tracking"

# ============================================================
# SEARCH — SPECIALIZED / VERTICAL
# ============================================================

[[search.specialized]]
name     = "arXiv"
domain   = "export.arxiv.org"
protocol = "https"
notes    = "Academic paper search and retrieval"

[[search.specialized]]
name     = "Semantic Scholar"
domain   = "api.semanticscholar.org"
protocol = "https"
notes    = "Research papers + citation graph"

[[search.specialized]]
name     = "PubMed / NCBI"
domain   = "eutils.ncbi.nlm.nih.gov"
protocol = "https"
notes    = "Biomedical literature"

[[search.specialized]]
name     = "Wikipedia API"
domain   = "en.wikipedia.org"
protocol = "https"
notes    = "Encyclopedic lookups"

[[search.specialized]]
name     = "Wikidata"
domain   = "www.wikidata.org"
protocol = "https"
notes    = "Structured knowledge graph queries"

[[search.specialized]]
name     = "NewsAPI"
domain   = "newsapi.org"
protocol = "https"
notes    = "Global news article aggregation"

[[search.specialized]]
name     = "GDELT"
domain   = "api.gdeltproject.org"
protocol = "https"
notes    = "Global event and news graph"

[[search.specialized]]
name     = "Hunter.io"
domain   = "api.hunter.io"
protocol = "https"
notes    = "Email and domain search"

[[search.specialized]]
name     = "BuiltWith"
domain   = "api.builtwith.com"
protocol = "https"
notes    = "Technology stack detection"

# ============================================================
# WEB CRAWL & CONTENT FETCH
# ============================================================

# Zero-setup URL-to-markdown; just prepend: r.jina.ai/<url>
[[web_crawl]]
name     = "Jina Reader"
domain   = "r.jina.ai"
protocol = "https"
notes    = "Prepend to any URL for instant clean markdown"

# URL-to-clean-Markdown; API pattern: defuddle.md/<url>
# MIT-licensed by @kepano; also available as NPM package for self-hosting
[[web_crawl]]
name     = "Defuddle"
domain   = "defuddle.md"
protocol = "https"
notes    = "Clean markdown extraction; same usage pattern as Jina Reader"

[[web_crawl]]
name     = "Firecrawl"
domain   = "api.firecrawl.dev"
protocol = "https"
notes    = "Deep crawl + structured markdown extraction"

[[web_crawl]]
name     = "Browserbase"
domain   = "api.browserbase.com"
protocol = "https"
notes    = "Headless browser cloud for JS-heavy pages"

[[web_crawl]]
name     = "Apify"
domain   = "api.apify.com"
protocol = "https"
notes    = "Scraper actors + cloud storage"

[[web_crawl]]
name     = "ScrapingBee"
domain   = "app.scrapingbee.com"
protocol = "https"
notes    = "Anti-bot bypass proxy"

[[web_crawl]]
name     = "Zyte"
domain   = "api.zyte.com"
protocol = "https"
notes    = "Enterprise-grade structured scraping"

# ============================================================
# MCP / TOOL SERVERS
# ============================================================

[[mcp_servers]]
name     = "Composio MCP"
domain   = "mcp.composio.dev"
protocol = "https"
notes    = "Multi-service MCP hub"

[[mcp_servers]]
name     = "Zapier MCP"
domain   = "mcp.zapier.com"
protocol = "https"
notes    = "Zapier automation via MCP"

[[mcp_servers]]
name     = "Asana MCP"
domain   = "mcp.asana.com"
protocol = "https"
notes    = "Task and project management"

[[mcp_servers]]
name     = "Linear"
domain   = "api.linear.app"
protocol = "https"
notes    = "Issue tracker MCP"

[[mcp_servers]]
name     = "Notion"
domain   = "api.notion.com"
protocol = "https"
notes    = "Pages and databases"

[[mcp_servers]]
name     = "Slack"
domain   = "api.slack.com"
protocol = "https"
notes    = "Slack messaging API; also slack.com for OAuth"

[[mcp_servers]]
name     = "Google Drive MCP"
domain   = "drivemcp.googleapis.com"
protocol = "https"
notes    = "Google Drive file access"

[[mcp_servers]]
name     = "Google Calendar MCP"
domain   = "calendarmcp.googleapis.com"
protocol = "https"
notes    = "Google Calendar events"

[[mcp_servers]]
name     = "Gmail MCP"
domain   = "gmailmcp.googleapis.com"
protocol = "https"
notes    = "Gmail read/send"

# ============================================================
# PACKAGE REGISTRIES
# ============================================================

[[package_registries]]
name     = "crates.io (index)"
domain   = "index.crates.io"
protocol = "https"
notes    = "Rust crate sparse index"

[[package_registries]]
name     = "crates.io (registry)"
domain   = "crates.io"
protocol = "https"
notes    = "Rust crate metadata"

[[package_registries]]
name     = "crates.io (static)"
domain   = "static.crates.io"
protocol = "https"
notes    = "Rust .crate file downloads"

[[package_registries]]
name     = "npm"
domain   = "registry.npmjs.org"
protocol = "https"
notes    = "Node.js packages"

[[package_registries]]
name     = "PyPI"
domain   = "pypi.org"
protocol = "https"
notes    = "Python package index"

[[package_registries]]
name     = "PyPI (files)"
domain   = "files.pythonhosted.org"
protocol = "https"
notes    = "Python package file downloads"

[[package_registries]]
name     = "Go Modules Proxy"
domain   = "proxy.golang.org"
protocol = "https"
notes    = "Go module proxy"

[[package_registries]]
name     = "Go Checksum DB"
domain   = "sum.golang.org"
protocol = "https"
notes    = "Go module checksum verification"

[[package_registries]]
name     = "Maven Central"
domain   = "repo1.maven.org"
protocol = "https"
notes    = "JVM artifact repository"

[[package_registries]]
name     = "RubyGems"
domain   = "rubygems.org"
protocol = "https"
notes    = "Ruby gem registry"

[[package_registries]]
name     = "NuGet"
domain   = "api.nuget.org"
protocol = "https"
notes    = ".NET package registry"

[[package_registries]]
name     = "GHCR"
domain   = "ghcr.io"
protocol = "https"
notes    = "GitHub Container Registry; Homebrew bottles + images"

# ============================================================
# CODE HOSTING & VERSION CONTROL
# ============================================================

[[code_hosting]]
name     = "GitHub API"
domain   = "api.github.com"
protocol = "https"
notes    = "REST + GraphQL API"

[[code_hosting]]
name     = "GitHub Raw"
domain   = "raw.githubusercontent.com"
protocol = "https"
notes    = "Raw file content from repos"

[[code_hosting]]
name     = "GitHub Code Downloads"
domain   = "codeload.github.com"
protocol = "https"
notes    = "Tarball/zip archive downloads"

[[code_hosting]]
name     = "GitHub Objects (LFS)"
domain   = "objects.githubusercontent.com"
protocol = "https"
notes    = "Git LFS and release asset storage"

[[code_hosting]]
name     = "GitHub"
domain   = "github.com"
protocol = "https"
notes    = "Web + OAuth + MCP server"

[[code_hosting]]
name     = "GitLab"
domain   = "gitlab.com"
protocol = "https"
notes    = "GitLab API + web"

[[code_hosting]]
name     = "Bitbucket"
domain   = "api.bitbucket.org"
protocol = "https"
notes    = "Bitbucket REST API"

# ============================================================
# AUTH & IDENTITY
# ============================================================

[[auth]]
name     = "GitHub OAuth"
domain   = "github.com"
path     = "/login/oauth"
protocol = "https"
notes    = "GitHub OAuth token exchange"

[[auth]]
name     = "Google OAuth"
domain   = "oauth2.googleapis.com"
protocol = "https"
notes    = "Google token exchange"

[[auth]]
name     = "Google Accounts"
domain   = "accounts.google.com"
protocol = "https"
notes    = "Google sign-in flow"

[[auth]]
name     = "Auth0"
domain   = "*.auth0.com"
protocol = "https"
# Wildcard — scope to your specific tenant subdomain in production
notes    = "Scope to tenant subdomain in prod"

[[auth]]
name     = "Clerk (accounts)"
domain   = "*.clerk.accounts.dev"
protocol = "https"
notes    = "Clerk hosted auth"

[[auth]]
name     = "Clerk API"
domain   = "api.clerk.com"
protocol = "https"
notes    = "Clerk backend SDK"

# ============================================================
# DEV INFRASTRUCTURE
# ============================================================

[[dev_infra]]
name     = "Cloudflare Workers"
domain   = "*.workers.dev"
protocol = "https"
notes    = "Serverless edge functions"

[[dev_infra]]
name     = "Vercel"
domain   = "vercel.com"
protocol = "https"
notes    = "Deployment platform + API"

[[dev_infra]]
name     = "Vercel Apps"
domain   = "*.vercel.app"
protocol = "https"
notes    = "Deployed app subdomains"

[[dev_infra]]
name     = "Render"
domain   = "*.onrender.com"
protocol = "https"
notes    = "Render hosted services"

[[dev_infra]]
name     = "Fly.io Machines"
domain   = "api.machines.dev"
protocol = "https"
notes    = "Fly.io Machines API"

[[dev_infra]]
name     = "Supabase"
domain   = "*.supabase.co"
protocol = "https"
notes    = "Postgres + auth + storage"

[[dev_infra]]
name     = "PlanetScale"
domain   = "*.psdb.cloud"
protocol = "https"
notes    = "Serverless MySQL"

[[dev_infra]]
name     = "Upstash"
domain   = "*.upstash.io"
protocol = "https"
notes    = "Serverless Redis + Kafka + Vector"

[[dev_infra]]
name     = "PostHog"
domain   = "us.posthog.com"
protocol = "https"
notes    = "Product analytics"

[[dev_infra]]
name     = "PostHog Ingestion"
domain   = "us.i.posthog.com"
protocol = "https"
notes    = "PostHog event ingestion endpoint"

# ============================================================
# OS / SYSTEM UPDATES (Ubuntu agent infra)
# ============================================================

[[os_updates]]
name     = "Ubuntu Security"
domain   = "security.ubuntu.com"
protocol = "http"
notes    = "Ubuntu security package mirror"

[[os_updates]]
name     = "Ubuntu Archive"
domain   = "archive.ubuntu.com"
protocol = "http"
notes    = "Ubuntu main package archive"