Poisoned Rust crates turned routine builds into credential theft for about 90 minutes each
Malicious versions of arrayref, internment, and append-only-vec shipped an infostealer targeting browser profiles and crypto wallets. crates.io pulled them within two hours, but lockfiles and registry caches keep them around.
Attackers published malicious releases of three widely used Rust crates on August 21, 2026: arrayref 0.3.10, with 245 million lifetime downloads, internment 0.8.7, and append-only-vec 0.1.9, all maintained by the same developer. A typosquatted proc-macro1 and several throwaway crates carried the payload, which targeted Chrome, Brave, and Edge browser profiles and cryptocurrency wallet extensions, established persistence, and accepted attacker commands on Linux, Windows, Intel Macs, and Apple Silicon. Exposure windows were 86, 90, and 107 minutes before crates.io removed them. The Rust team suspects the maintainer's machine or credentials were compromised but has not stated how.
Why it matters
A 90-minute window is not a small one when CI runs on every push. Any pipeline that resolved a fresh dependency graph in that period pulled the payload, and unlike a developer laptop, a build runner holds registry tokens, cloud credentials, and signing keys. The compromise path here is the recurring one in package ecosystems: the maintainer account rather than the code. Publishing credentials are long-lived, often unrotated, and frequently protected by nothing stronger than a password. Concrete steps: pin and commit Cargo.lock, check lockfiles and local registry caches for the affected versions, require phishing-resistant MFA on publish accounts, and scope publish tokens to single crates. Rotate anything a build runner touched during the window.
Source: The Register
Related on Start with Identity
- BlogFortiBleed: a firewall packet capture turned into a credential harvester
An initial access broker abused FortiOS's own packet-capture feature with a Go tool called FortigateSniffer, reading cleartext passwords and Kerberos and NTLM h
- BlogGitGuardian found 4,576 leaked n8n tokens, and a third of reachable instances accepted them
Exposed n8n API tokens in public GitHub commits gave researchers read access to workflows, stored credentials, and data tables, plus a path to exfiltrate the ra
- BlogOver 1,000 AI-named npm typosquats deliver a cross-platform dropper
OpenSourceMalware and Sonatype tracked Flooding Dropper, a campaign of AI-generated typosquat package names carrying WEL1DROPPER, which fetches Sliver on Linux
- RankingBest Verifiable Credential Platforms: Issuance & Verification Infrastructure
The best verifiable credential platforms in 2026: MATTR, Procivis, SpruceID, walt.id, and Hyperledger. Ranked for standards conformance (W3C VC, SD-JWT, OpenID4
- ArticleCIAM vs IAM: Key Differences and When You Need Each
CIAM and IAM both manage identity, but they solve opposite problems. This guide explains the differences in users, scale, priorities, and architecture, and how
- GlossaryCredential Stuffing
An attack that replays username and password pairs leaked from other breaches against a target, exploiting password reuse. Defended with MFA, passkeys, and bot