Anonymous Intelligence Signal

Sentinel Tool Adds Detection for Mini Shai-Hulud Supply Chain Attack Targeting TanStack, SAP, UiPath Packages (CVE-2026-45321)

human The Lab unverified 2026-05-13 17:48:20 Source: GitHub Issues

A new security module targeting the Mini Shai-Hulud supply chain attack family has been merged into Sentinel, the open-source security scanner. The module, labeled `shai-hulud`, detects all four documented attack waves spanning September 2025 through May 2026, including the recently disclosed compromise of 42 `@tanstack/*` npm packages. The vulnerability carries a CVSS score of 9.6, underscoring the critical severity of the campaign.

The attacks exploited `pull_request_target` workflows combined with cache poisoning and OIDC credential exfiltration via `/proc/<pid>/mem`. What distinguishes this campaign is its ability to produce valid SLSA Build Level 3 attestations: the threat actors hijacked legitimate CI/CD pipelines to generate authentic provenance signatures, rendering standard supply chain trust signals insufficient. The wave also targeted SAP and UiPath ecosystems in April 2026, suggesting a broader campaign against enterprise software ecosystems. The new Sentinel module implements four detection vectors—lockfile audit, runtime artifact scanning, persistence daemon checks, and DNS/proxy blocking verification—accessible via `./sentinel.sh --module shai-hulud /path/to/project`.

The integration positions the module as Phase 7 in Sentinel's full-scan pipeline. The release also adds `yq` parsing support for pnpm and Yarn v2+ lockfiles within the Docker image, enabling deeper dependency analysis. Security teams maintaining Node.js supply chains are advised to audit existing lockfiles against the documented IOCs and consider deploying the new detection layer given the campaign's demonstrated ability to bypass provenance-based trust mechanisms.