Malware and known malicious packages
Detects malicious payloads, suspicious install hooks and packages flagged in malware databases.
Bytesafe sits between developers, CI/CD pipelines and public package registries to inspect package requests, enforce policy and stop malicious, vulnerable or non-compliant dependencies before they enter your environment.
Bytesafe intercepts every package request before the package reaches a developer, pipeline, or internal repository.
Public registries
Bytesafe Dependency Firewall
Policy engine
Your environment
Public registries
Bytesafe Dependency Firewall
Policy engine
Your environment
From the moment a package is requested to the moment it is allowed, blocked or flagged for review.
A developer, CI/CD pipeline or AI agent runs a package install.
Bytesafe receives the request before the package reaches the environment.
The package is checked against policy: malware, CVEs, licenses, age and provenance.
The package is approved or blocked. Exceptions can be approved by an admin.
The result is recorded with the rule, requester and timestamp.
Every package request is evaluated in real time against the active policy set.
Detects malicious payloads, suspicious install hooks and packages flagged in malware databases.
Blocks packages with CVEs above a configurable CVSS or EPSS score. New advisories take effect immediately.
Blocks public packages that impersonate internal ones. Private packages always resolve first.
Blocks packages with disallowed licenses before they reach a build.
Holds newly published versions for a configurable window (7 or 14 days) to let the ecosystem surface zero-day issues.
Verifies package origin using Sigstore and SLSA attestations. Detects version swaps and pipeline tampering.
Supports npm, PyPI, Maven, NuGet, Go and OCI with ecosystem-aware policy controls.
Block or allowlist by package name, version range, requester, team or upstream source.
Each rule targets one or more conditions. Rules either block the request or log it. Multiple rules stack per firewall and take effect immediately on save.
What the policy engine can filter on
Secure defaults (always-on rules)
Rules as code
Every package request ends in one of four states: approved, blocked, delayed or approved by exception. All are recorded with the same level of detail.
The package meets policy. It passes through and reaches the developer or pipeline with no interruption.
The package violates policy. It is stopped before install. The developer and CI/CD log show the rule that triggered.
The package version is newly published. It is held for a configurable window (7 or 14 days) to give the ecosystem time to surface zero-day issues before it reaches your environment.
The package would normally be blocked, but an admin has approved an exception. The exception is time-limited and logged with the reason.
Bytesafe acts as a proxy in front of your existing repository. Developers and CI/CD pipelines keep their existing registry URLs and credentials. No migration required.
Supported ecosystems
Scanners find issues after dependencies are already in the environment. A dependency firewall stops them before they arrive.
Scanner-based detection
Firewall-based prevention
Security, DevOps and engineering teams use Dependency Firewall to stop risky packages and enforce policy across their development environment.
Prevent known malware and account-takeover packages from reaching any developer or build.
Learn moreRoute pipelines through the firewall and apply the same rules on every build, automatically.
Learn moreEnsure internal package names always resolve from your private registry.
Learn moreApply different policies per team, project, or pipeline with separate firewall configurations.
Learn moreAdmins can grant exceptions for packages that would otherwise be blocked, with a reason and expiry logged.
Learn moreEvery allow, block and exception is logged with the rule, requester and timestamp.
Learn moreSee how Bytesafe can protect your developers, builds and existing repository setup from risky open source packages.