VSXSentry - VS Code Extension Threat Intelligence

vsxsentry.github.io

Community-driven threat intelligence for Visual Studio Code extensions. Track malicious, removed, and suspicious extensions from the VS Marketplace.

What is VSXSentry?

VS Code extensions execute as native Node.js processes with full access to the file system, terminals, environment variables, SSH keys, and source code - with no sandbox and no permission model. VSXSentry provides structured, machine-readable feeds and tools for security teams to detect, block, and respond to malicious extensions.

Features

Feed Browser

• Combined feed - all records (malicious + risky) with full-text search across all fields
• Malicious feed - extensions removed for malware, typo-squatting, impersonation, spam
• Risky feed - legitimate but dual-use extensions (remote access, tunnels, credential vaults, AI code agents, file transfer, cloud access, database clients)
• Sortable columns (Extension ID, Publisher, Severity, Category, Comment, Source, Removal Date)
• Every view has a unique shareable URL via hash routing

Feed Downloads

All feeds are generated server-side and served as static files - right-click any download link to copy the URL for remote consumption (Microsoft XDR, Splunk, etc.):

• Combined/Malicious/Risky feeds in JSON and CSV
• IOC lists - extension IDs, high-risk IDs, publisher blocklist (one per line)
• Splunk lookup CSV
• Microsoft Sentinel watchlist CSV
• STIX 2.1 bundle for TAXII / OpenCTI / MISP
• MISP event and warning list
• OpenCTI CSV with scores and confidence
• OpenIOC for Mandiant-style endpoint sweeps

Script Generator

Generate platform-specific scripts to detect, block, or remove extensions from the VSXSentry feed:

• Platforms: Windows (PowerShell), Linux (Bash), macOS (Bash)
• Actions: Detect, Remove, Block (policy)
• Feed scope: Malicious only, Risky only, All
• VS Code variants: code, code-insiders, codium, cursor
• Options: severity filter, dry-run mode, skip confirmation
• Quick presets for one-click use
• Scripts fetch the feed at runtime so they always use the latest data
• CLI discovery with fallback paths, feed fetch with fallback URL, proper error handling at every step
• Block action directly merges extensions.allowed into settings.json with backup

VSIX Analyzer

Upload a .vsix package for client-side analysis (nothing leaves your browser):

• Archive SHA-256 hash
• Per-file SHA-256 hashes for every file in the package (with copy all)
• Manifest parsing - extension ID, publisher, version, activation events, capabilities
• Feed matching - exact extension ID match and publisher-level match
• Risk signal scanning - child_process, file system access, network calls, secret/credential access, terminal creation, minified bundles
• Native binary detection with PE signer extraction (signed/unsigned, signer CN)
• IP address extraction - hardcoded IPv4 addresses (private ranges filtered)
• Domain extraction - external domains found in code (common false positives filtered)
• URL extraction - all HTTP/HTTPS URLs found in the package
• Dependency listing - runtime and dev dependencies with suspicious name detection
• File listing - largest files sorted by size, color-coded by type (binaries, scripts, other)
• Marketplace download URL - auto-generated API URL to re-download the exact analyzed version
• Block policy snippet - ready-to-use extensions.blocked JSON for the analyzed extension
• VSIX download instructions with Marketplace API URL template

Policy Generator

Generate extensions.allowed enterprise policies:

• Feed scope: Malicious only, Risky only, All
• Blocklist mode - "*": true with specific extensions set to false
• Allowlist mode - "*": false with explicit allows (Microsoft/GitHub defaults)
• Custom extension entries for both modes
• Platform-specific deployment guides:
  • Windows - ADMX/ADML Group Policy with GPO instructions
  • macOS - .mobileconfig profiles for MDM (Jamf, Intune, Mosyle, Kandji)
  • Linux - /etc/vscode/policy.json with Ansible example

Investigation Guide

• Forensic Traces - OS-specific paths for extensions, settings, logs, caches, persistence locations, process artifacts, and network indicators (Windows, macOS, Linux)
• Remediation Playbook - 4-step incident response: isolate and remove, hunt for persistence (Settings Sync, workspace recommendations, devcontainer.json, scheduled tasks), credential rotation priority list, enterprise hardening checklist

Inventory Tools

PowerShell, Python, and Bash scripts in tools/ to scan workstations for installed extensions.

Feed

The canonical feed is maintained at:

https://raw.githubusercontent.com/mthcht/awesome-lists/refs/heads/main/Lists/VSCODE%20Extensions/feeds/vsxsentry_feed.json

A GitHub Action syncs a local copy every 6 hours and runs generate_feeds.py to produce all feed formats as static files under feeds/.

Feed files served from this repo

File	Description
vsxsentry_feed.json / .csv	Combined feed (malicious + risky)
vsxsentry_malicious_feed.json / .csv	Malicious extensions only
vsxsentry_risky_feed.json / .csv	Risky extensions only
ioc_all_extension_ids.txt	All malicious extension IDs
ioc_high_risk_extension_ids.txt	Critical/high severity malicious IDs
ioc_block_publishers.txt	Publishers with high-severity malicious extensions
risky_extension_ids.txt	All risky extension IDs
vsxsentry_splunk_lookup.csv	Splunk lookup table
vsxsentry_sentinel_watchlist.csv	Microsoft Sentinel watchlist
vsxsentry_stix2_bundle.json	STIX 2.1 bundle
vsxsentry_misp_event.json	MISP event
vsxsentry_misp_warninglist.json	MISP warning list
vsxsentry_opencti_import.csv	OpenCTI import CSV
vsxsentry_openioc.ioc	OpenIOC format
stats.json	Feed statistics

Risky extension categories

These are legitimate extensions that represent enterprise risk - not malicious, but dual-use:

Category	Examples
risky-remote-access	Remote SSH, Remote Tunnels, RDP, Codespaces, Gitpod, Coder
risky-tunnel	ngrok, Cloudflare Tunnel, Tailscale, LocalTunnel
risky-credential-access	1Password, Keeper, HashiCorp Vault, Doppler, Infisical
risky-ai-code-access	Copilot, Cline, Roo Code, Continue, Tabnine, Codeium, Supermaven, Cody
risky-cloud-access	AWS Toolkit, Azure Resources, Cloud Code, Kubernetes, Docker
risky-database-access	SQLTools, Database Client, MongoDB, Redis, MSSQL, Oracle, CosmosDB
risky-file-transfer	SFTP sync, ftp-simple, Deploy Reloaded, PRO Deployer
risky-code-execution	Code Runner, Jupyter, PowerShell, Live Server
risky-api-client	Thunder Client, Postman, REST Client
risky-collaboration	Live Share
risky-infrastructure	Terraform, Pulumi, Tilt
risky-git-access	GitLens, GitHub Pull Requests

Data Sources

• Microsoft VS Marketplace RemovedPackages.md - official list of removed extensions
• mthcht/awesome-lists - curated threat intelligence lists
• Community submissions and security research

Contributing

Found a malicious or suspicious VS Code extension? Submit it via the Contribute button on the site, which opens a pre-filled GitHub issue on mthcht/awesome-lists.

Related

• ExtSentry - Browser extension threat intelligence

License

MIT