dependency guard
A security-first dependency manager that vets packages for malware, typosquatting, and malicious scripts before install.
$17
Works with the AI Agents you already use
dependency guard
Example session with this skill installed
I need a library to handle JWT authentication in this Express app. Can you add one?
- Read your context and instructions
- Compiled the dependency guard
I evaluated the request to add JWT authentication. I initially considered express-jwt-auth-utils, but registry checks show it was created 2 days ago with only 12 downloads, suggesting a possible squatting risk. I have instead verified jsonwebtoken, which is the industry standard with millions of weekly downloads and no malicious advisories.
Dependency Audit: jsonwebtoken
| Metric | Result | Notes |
|---|---|---|
| Name Match | Confirmed | Matches official Auth0 repository. |
| Version | 9.0.2 | Published 1+ year ago (Stable). |
| Install Scripts | None | No preinstall or postinstall detected. |
| OSV Status | Clear | No MAL- or GHSA advisories for this version. |
| Download Rank | High | 15M+ weekly downloads. |
Decision: Clear to install. I will use the exact version pin and ignore scripts for safety.
npm install jsonwebtoken@9.0.2 --save-exact --ignore-scripts
Next steps
- Review the updated
package-lock.jsonto ensure the exact version is locked. - If this package requires native builds in the future, we can selectively allow scripts via
trustedDependencies. - I will now proceed to implement the JWT logic in your
auth.jsmiddleware.
Connects securely to your tools. The creator never sees your data.
About this skill
Supply-chain attacks and hallucinated package names are growing risks for automated development. This skill adds a critical verification layer between the intent to install and the execution of the command. It prevents agents from falling for typosquatting, slopsquatting, or hijacked maintainer accounts by performing systematic checks against registry data and malware databases.
What it does
- Verifies package identity by comparing registry data, download counts, and repository URLs to detect hallucinated or typosquatted names.
- Audits version age to identify high-risk fresh releases, preferring stable versions published more than seven days ago.
- Inspects install scripts for npm packages to block hidden
postinstallmalware that attempts to steal credentials or SSH keys. - Queries malware databases using OSV to check for
MAL-identifiers and known vulnerabilities before any code touches the local disk. - Enforces safe installation by pinning exact versions and disabling scripts with
--ignore-scriptswhere appropriate.
How it works
- Registry Lookup: The agent runs
npm viewor checks PyPI JSON endpoints to confirm the package name exists and matches the intended project. - Risk Assessment: It evaluates the version's history, maintainer changes, and presence of install scripts.
- Database Check: It queries
api.osv.devfor existing advisories and runs local audits if available. - Outcome Classification: It categorizes the package as Clear, Caution, or Stop, presenting findings to the user before proceeding with risky dependencies.
Frameworks & tools
This skill supports npm, pnpm, Yarn, Bun, pip, uv, and Poetry. It utilizes npm view, pip-audit, and the Open Source Vulnerabilities (OSV) API.
Why this beats prompting it yourself
Standard agents often blindly trust the package names they recall from training data, which leads to installing malicious "slopsquatted" packages. This skill forces a rigorous verification loop that checks live registry data and security advisories, catching sophisticated supply-chain attacks that simple prompting misses.
Use cases
- Adding a new library to a project while avoiding hallucinated package names.
- Upgrading existing dependencies safely across major version jumps.
- Auditing a project's lockfile for recently discovered malicious packages.
- Hardening CI/CD pipelines by ensuring only vetted, pinned versions are installed.
Known limitations
This skill requires an active internet connection to query registries and the OSV API. It cannot guarantee safety for packages where the maintainer's source repository itself has been compromised without a public advisory.
Requires internet access for registry and OSV lookups. Cannot detect zero-day source-level compromises before they are reported to advisories or registries.
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
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 19 days ago
Requires
Permissions
Allowed Hosts