Rosa Del Mar

Daily Brief

Issue 43 2026-02-12

Registry Operator Constraints: Resourcing, Economics, And Incentive Alignment

Issue 43 Edition 2026-02-12 9 min read
General
Sources: 1 • Confidence: Medium • Updated: 2026-02-12 18:47

Key takeaways

  • The npm team is perceived as under-resourced with a large backlog, contributing to slow responses unless issues become emergencies.
  • Nicholas argued that a new registry run by an AI company would face major trust and operational barriers, and separately argued that reliability issues in hosted AI services would make an AI-native npm-compatible registry less compelling.
  • About 500 npm packages were compromised in September 2025 alone.
  • Trusted publishing currently lacks two-factor authentication, and OpenJS Foundation guidance recommends critical packages avoid it due to repo-takeover publishing risk.
  • Recent malicious npm packages have targeted crypto theft and secret harvesting, including running TruffleHog to exfiltrate local secrets and propagate.

Sections

Registry Operator Constraints: Resourcing, Economics, And Incentive Alignment

  • The npm team is perceived as under-resourced with a large backlog, contributing to slow responses unless issues become emergencies.
  • Nicholas believes GitHub spinning npm off into a foundation is unlikely, and outlined a possible off-ramp involving joint funding via OpenJS to hire registry engineers.
  • Running the npm registry is extremely expensive and was a key reason npm Inc. needed to sell, making the registry a major cost center for its owner.
  • Nicholas claims npm's main differentiator versus other package ecosystems is scale, which makes many solutions harder to apply to npm.
  • Nicholas explained that VC funding creates pressure to generate returns, and failure to monetize can lead investors to push for a sale.
  • Nicholas expects npm is capable of real-time package upload analysis and signature-based blocking, but believes this capability is not being applied broadly enough to prevent repeated similar attacks.

Competitive Landscape: Alternative Registries And Client-Layer Mitigations

  • Nicholas argued that a new registry run by an AI company would face major trust and operational barriers, and separately argued that reliability issues in hosted AI services would make an AI-native npm-compatible registry less compelling.
  • JSR launched with reserved/confusing scopes, trusted publishing and 2FA requirements, and no pre/post-install scripts.
  • pnpm shipped a client-side mitigation that blocks installing packages newer than about seven days to slow rapid spread of compromised releases.
  • JSR's early rapid iteration and issue responsiveness degraded over time to the point that bug reports stopped getting responses.
  • JSR's announced open-governance committee produced no visible progress or updates and effectively went nowhere.
  • JSR did not support a workable way to publish npm packages that depend on JSR packages, leading at least one ESLint package to copy JSR source code into its own repository.

Npm Supply-Chain Compromise Magnitude And Attacker Mechanics

  • About 500 npm packages were compromised in September 2025 alone.
  • ESLint maintainers periodically receive suspicious pull requests that only change dependencies without explanation.
  • ESLint is downloaded at a rate of over 200 million times per month.
  • A common npm compromise pathway is stolen maintainer credentials followed by publishing a malicious package that uses preinstall/postinstall scripts to execute on install.
  • Recent malicious npm packages have targeted crypto theft and secret harvesting, including running TruffleHog to exfiltrate local secrets and propagate.
  • ESLint previously experienced an npm compromise enabled by a maintainer reusing credentials from another site that was hacked, which allowed unauthorized publishing.

Publisher Authentication Changes: Token Rotation Vs Oidc Trusted Publishing

  • Trusted publishing currently lacks two-factor authentication, and OpenJS Foundation guidance recommends critical packages avoid it due to repo-takeover publishing risk.
  • npm token policy changes converged on fine-grained tokens with a limited lifetime of about 90 days, requiring periodic rotation or automation.
  • Scaling token and publishing changes was difficult for maintainers with many packages because npm initially lacked batch operations and required per-package setup with repeated 2FA prompts.
  • Trusted publishing uses OIDC from CI (e.g., GitHub Actions) to mint short-lived npm publish tokens on demand instead of storing long-lived tokens.
  • Trusted publishing is enforced by pre-registering a specific repository and a specific workflow file as the only allowed publishing source for a package.
  • GitHub's npm security changes shift more responsibility and operational burden onto maintainers rather than addressing root causes at the registry level.

Install Scripts As An Ecosystem-Level Attack Surface With Constrained Mitigation Space

  • Recent malicious npm packages have targeted crypto theft and secret harvesting, including running TruffleHog to exfiltrate local secrets and propagate.
  • npm pre- and post-install scripts can execute arbitrary commands, creating a significant attack surface in a public registry context.
  • Banning pre/post-install scripts on npm would break a non-trivial share of packages because native modules rely on those scripts to compile per-machine artifacts.
  • Using npm's ignore-scripts option reduces risk but can silently break dependency trees that require compilation, making failures hard to debug.
  • Nicholas proposed requiring a major version bump when a package adds install scripts to prevent automatic uptake via patch/minor updates.
  • A proposed mitigation is heightened automated scrutiny or a waiting period for publishes that add or change install scripts.

Watchlist

  • ESLint maintainers periodically receive suspicious pull requests that only change dependencies without explanation.
  • The npm team is perceived as under-resourced with a large backlog, contributing to slow responses unless issues become emergencies.
  • A gap was identified in maintainer-focused resources for leveling up security/operations, with a possible lead that OpenJS may have produced something similar.
  • A single major npm supply-chain attack could plausibly cause multi-million-dollar damage and trigger serious reputational and possibly legal consequences for the registry operator.
  • Jerod suggests that because JSR was intended to be open and governed, another party could potentially pick it up if the Deno team stops progressing it.

Unknowns

  • What is the precise source, definition, and verification method for the 'about 500 compromised npm packages in September 2025' figure?
  • How often do compromises involve account takeover versus malicious PR/contribution routes, and what share of incidents are driven by install scripts specifically?
  • What are adoption rates of fine-grained expiring tokens and OIDC trusted publishing across high-impact npm packages, and how does adoption correlate with reduced incidents?
  • What specific OpenJS guidance exists on trusted publishing for critical packages, and what changes would be required (e.g., step-up verification) for that guidance to change?
  • Does npm actually perform real-time upload analysis and signature-based blocking at scale today, and what policies determine when it is applied?

Investor overlay

Read-throughs

  • Rising npm supply chain incident volume and install script abuse could increase demand for developer focused supply chain security tools that detect malicious dependency changes and secret exfiltration behaviors.
  • Perceived under resourcing and backlog at the registry operator could create opportunity for third party services that add trust and operational controls without requiring a new registry.
  • Stalled progress of alternative registries with stronger defaults suggests client layer mitigations may be the most viable near term path, potentially benefiting tooling that enforces freshness windows and friction on risky changes.

What would confirm

  • Credible reporting that compromised npm packages remain frequent and that attacks commonly use install scripts and credential theft, alongside increasing downstream damage estimates.
  • Meaningful adoption increases of expiring fine grained tokens or OIDC trusted publishing across high impact packages, and evidence that adoption correlates with fewer incidents.
  • Evidence that registries do not perform upload analysis at scale or apply it narrowly, paired with continued slow response unless emergencies, pushing enterprises toward external controls.

What would kill

  • Verification that the cited compromised package figure is inaccurate or that incident volume is materially lower than implied, reducing urgency and budgets.
  • Registry operator implements effective real time upload analysis and blocking at scale with clear policies, reducing the need for third party overlays.
  • Trusted publishing gains strong step up verification or 2FA and becomes broadly recommended for critical packages, materially lowering repo takeover publishing risk.

Sources