npm-safe: Scan npm Packages Before You Install
It invokes the open-source `@npm-safe/core` CLI on your machine and explains the actual findings, scan boundaries and next steps.
Free
Works with the AI Agents you already use
About this skill
Scan before you install
npm-safe helps an AI coding agent inspect npm packages for supply-chain risk signals before a dependency change. It invokes the open-source @npm-safe/core CLI on your machine and explains the actual findings, scan boundaries and next steps.
What it does
- Checks published package metadata and README signals for suspicious scripts, naming risks and other supply-chain indicators.
- Supports opt-in
--deepchecks that verify and statically inspect the published tarball within resource limits, without extracting or executing the inspected package code. - Reports an explainable 0-100 signal score, severity and actionable findings; a clean result is not a guarantee of safety.
- Scans direct dependencies or exact versions from
package-lock.jsonfor CI workflows. - Supports batch checks, local watchlists and JSON/CSV report exports when requested.
How it works
Your agent checks that npm-safe is available, selects the scan appropriate to your request, runs it locally, and explains the result. Dependency installation and configuration changes stay within your instructions and existing authorization.
Requirements
Node.js 20.12.0 or newer and @npm-safe/core 2.0.0 or newer must be installed separately. Your agent needs terminal access, access to the configured npm registry, and a writable local scanner cache. A chat-only agent without a terminal cannot run the scanner.
Data and limitations
Static scanning and caching run locally, with registry requests. Optional LLM analysis is not required: if enabled by the user, it can send metadata or selected source to the configured model provider and may incur that provider's charges. Deep scans have resource limits and can be incomplete. This skill does not install a persistent hook or guarantee interception of every installation. It complements dependency review and tools such as npm audit; it does not prove runtime safety.
Free and open source
This is a free skill for the Apache-2.0-licensed npm-safe project. The CLI is a separate open-source prerequisite. Source, setup instructions and issue reporting are available at https://github.com/nisconder/npm-safe. The archive includes the original license and a record of its small Agensi delivery adaptation.
A scan is evidence within a defined boundary, not proof that a package is safe. Default checks inspect metadata and README; tarball inspection requires --deep. Resource limits, unavailable metadata, unsupported lockfile entries or network failures may produce incomplete or unknown results. The local CLI must be installed separately. No persistent install hook is provided. Optional model analysis can transmit metadata/source and incur provider charges.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before any payment.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Free forever
No account required to browse
Trust & safety
Security scan
Security scan passed 4 days ago
- Free with an Agensi account
Requires
Permissions
Runs npm-safe through the terminal and contacts the configured npm registry. Reads project manifests/lockfiles for audits. Writes scanner cache/check history locally and report files when requested. Scanning does not execute inspected package code. Optional LLM analysis contacts the configured provider only under user authorization.