supply-chain
Safeguard articles tagged "supply-chain" — guides, analysis, and best practices for software supply chain and application security.
859 articles
Homomorphic Encryption in Software Supply Chains
A grounded look at BFV, CKKS, and TFHE schemes for supply chain workloads, measured costs, library choices, and where HE is not yet practical.
What an SBOM Scanner Does and How to Choose One
An SBOM scanner reads a software bill of materials and matches every listed component against vulnerability data. Here is how that differs from source scanning and what makes one worth trusting.
Is the mammoth npm Package Safe? A DOCX Converter Security Review
The mammoth npm package converts .docx files to HTML, but CVE-2025-11849 showed how a crafted document can read files off your server. Here is what to check.
Every Supply Chain Attack of June and July 2026 Was After the Same Thing
Nine incidents in eight weeks: a PyPI worm, typosquatted payment SDKs, jscrambler, AsyncAPI, Hugging Face, Polymarket, Nx Console, Medtronic, AdaptHealth. Nine different vectors, one prize — credentials sitting in developer environments and build pipelines. If you fix one thing this quarter, fix that.
Eighteen Minutes: The Nx Console Extension Compromise and the IDE Blind Spot
A poisoned VS Code extension was live for eighteen minutes. In that window, auto-update pushed it into every developer environment with Nx Console installed — including a GitHub employee's device, leading to exfiltration of internal GitHub repositories. Your IDE extensions have no SBOM, no review, and a direct push channel to your engineers.
Polymarket Lost ~$3M Without a Single Smart Contract Bug
On 25–26 June 2026 attackers compromised a third-party vendor and injected malicious code into Polymarket's website frontend, manipulating users into approving fraudulent transactions. Roughly $3M in crypto drained. The smart contracts were never touched. Your client-side dependency tree is production.
Python's .pth Files Are a Code Execution Primitive, and Attackers Noticed
The June 2026 PyPI worm wave used a *-setup.pth file to execute at interpreter startup — before your code, before your imports, on every single python invocation. It then fetched the Bun JavaScript runtime to run its payload. If your supply chain model stops at setup.py, it has a hole in it.
Seventeen Fake Payment SDKs, Six Minutes to Detection, and a Sandbox Check
On 7 July 2026, roughly 17 typosquatted payment-provider packages hit npm and PyPI — paysafe-checkout, paysafe-node, neteller and friends. They swept environment variables matching KEY, SECRET, TOKEN, PASS, AUTH and API, explicitly hunted AWS_SECRET_ACCESS_KEY, GITHUB_TOKEN and NPM_TOKEN, and exited quietly if they thought they were in a sandbox.
npm v12 Disabled Install Scripts. Attackers Adapted in Three Days.
On 8 July 2026 npm v12 shipped with install scripts off by default — closing what GitHub called the ecosystem's largest code-execution surface. By 11 July, the jscrambler payload was executing on import instead. A study in why one-vector mitigations buy days, not years.
The AsyncAPI Hijack: When Trusted Publishing Becomes the Attack Path
On 14 July 2026 attackers used 37 pull requests against a pull_request_target workflow to steal the asyncapi-bot token, then let npm's OIDC trusted publisher automatically ship the malicious release. Four packages, 2.25 million weekly downloads, four hours live — and no code review was bypassed, because none was required.
The Jscrambler npm Compromise Went After Your AI Coding Assistant's Credentials
On 11 July 2026, five versions of the jscrambler package plus its webpack, gulp, grunt and metro plugins shipped malicious native binaries. The payload targeted crypto wallets and the credential stores of Claude Desktop, Cursor and Windsurf — and later versions fired on import, not install, defeating --ignore-scripts.
The Hugging Face Breach: What Changes When an AI Agent Runs the Intrusion
On 16 July 2026 Hugging Face disclosed that a malicious dataset gave an attacker code execution inside its data-processing pipeline, escalating to node-level access and internal cluster credentials over a single weekend — driven by an autonomous agent framework executing thousands of actions. Here's the anatomy, and what it means for anyone who treats a model registry as a trusted input.
Self-healing security runs on Safeguard.
Your first fix PR is minutes away.
No sales call required, even your agent can complete the purchase over MCP.