Launched January 6, 2026
vet

vet

Protect against Malicious Open Source Packages

FreeCLISecurityDevOps219 impressions#35 of its week

vet is a tool for protecting against open source software supply chain attacks. To adapt to organisational needs, it uses an opinionated policy expressed as Common Expressions Language (CEL) and extensive package security metadata including: - Code Analysis to guard against risks that actually matter - Vulnerabilities from OSV - Popularity based guardrails to prevent unvetted or risky packages - Maintenance status of the package - Extended License Attributes based compliance - OpenSSF Scorecard based 3rd party OSS risk management - Direct and Transitive dependency analysis for coverage

SafeDep blocks malicious open-source packages at install time across dev machines, CI/CD, repos and AI agents.

for
Developers, security teams and organizations that ship code using third-party packages.
pricing
freemium · free trial
license
Apache-2.0
safedep/vet 1.1k 113Goupdated 24 days ago
works withnpmpnpmpipuvbunGitHubClaude CodeCursorWindsurf

Key features

6 features of vet
  • Install-time blocking — Intercepts npm, pnpm, pip, uv, bun and other managers to refuse malicious packages before they run.
  • Policy-as-code enforcement — One CEL policy applied at every SDLC endpoint – dev machine, CI, repo and AI agent.
  • Org-wide inventory — Tracks packages, AI agents, skills and MCP servers per machine, project and user.
  • Pull-request checks — GitHub App posts malware, vulnerability and license verdicts on each PR as a required check.
  • CI/CD proxy — Runs as a proxy in pipelines, stopping malicious installs and failing the job on violations.
  • Threat intel feed — Optional add-on provides a continuous stream of confirmed malicious packages and indicators.

Use cases

  • Block malicious npm or PyPI packages on developers' laptops.
  • Prevent AI coding assistants from pulling poisoned dependencies.
  • Enforce package safety in CI pipelines before build artifacts are created.
  • Require clean dependency scans on every pull request merge.
  • Free$0 foreverUp to 3 SDLC endpoints · Real-time malicious package protection · 7-day findings history · Centralized policies
  • Team$80 per month/monthReal-time protection across all endpoints · 90-day findings history & audit trail · 50 on-demand package scans / month · Email support

vet vs alternatives

vetSnykPulumiNeovimDoppler
Best forPolicy-driven supply-chain protectionVulnerability managementIaC deploymentAI pair programmingSecret storage
PricingFreemiumSubscriptionFreeFreeSubscription
DevHunt upvotes00000
LaunchedJan 2026————
  • vet vs Snyk: Focuses on vulnerability scanning of code and containers, not real-time package install blocking
  • vet vs Pulumi: Infrastructure-as-code provisioning, not supply-chain package protection
  • vet vs Neovim: Open-source AI coding agent, whereas SafeDep secures packages for agents
  • vet vs Doppler: Secrets management, not package safety enforcement

vet FAQ

Do I need a credit card to start?+

No, the 30-day trial requires only basic details and no card.

How are CI runners counted as endpoints?+

Ephemeral runners are fingerprinted to a single CI endpoint, not counted per build.

What happens after the trial ends?+

You automatically move to the Free plan with protection on up to 3 endpoints; org-wide features pause until you subscribe.

Are the open-source tools still free?+

Yes, vet, pmg and gryph remain free, open-source and unmetered with no account required.

How does SafeDep differ from EDR?+

EDR monitors processes after they run; SafeDep blocks malicious packages before they are installed in the SDLC.

Summarized by DevHunt from safedep.io · Oct 6, 2026. Details may change; check the official site.

vet in the news

Recent vet news and updates from DevHunt Daily.

  • · Hacker News

    LiteLLM 1.107.0 on PyPI contained a credential stealer and backdoor

    The 1.107.0 release of the LiteLLM Python package was found to install malicious files that steal SSH keys, cloud credentials and other secrets, and run a remote code execution daemon. The malicious code was removed from PyPI after discovery, highlighting the risk of compromised packages for developers.