Google confirmed that eight emSign intermediate CA certificates had keys generated with insufficient randomness. The article gives Python and CI/CD users update and verification steps; exploitation against non-browser clients remains unknown.
StepSecurity found credential-stealing and remote-code-loader payloads in LiteLLM 1.107.0 PyPI wheels and Docker images. The stealer targeted developer, cloud and AI credentials; BerriAI removed affected artifacts and advised treating hosts as compromised.
Step Security reports OVHcloud CLI assets for five releases were swapped for a credential stealer in a 94-minute window. Its replay reproduced build fingerprints and loader behavior, identified installation-path differences, and outlines credential-rotation steps.
OpenCode says an ongoing breach affected 1.4% of Console users with linked Anthropic, OpenAI, or Google providers; the cause and accessed data remain undisclosed. StepSecurity advises inventorying installations, checking credentials, rotating keys, and reviewing usage.
Step Security reports that an Oct. 8 GhostAction wave injected malicious GitHub Actions workflows into 345 repositories, harvesting configured secrets and credentials from working trees and full Git histories. A run log confirmed exfiltration from uber/athenadriver.
StepSecurity says [email protected] steals credentials, spreads through npm packages and GitHub repositories, and can delete a user's home directory if its stolen GitHub token is revoked while a monitor is active.
StepSecurity introduced inventories of AI agent skills on developer machines and in GitHub repositories, showing execution-related flags, sources and content differences to help teams review skills that may run code or hooks.
StepSecurity reports that @subql/[email protected] contains a payload that runs on install or import, collects credentials, and supports exfiltration and remote commands via ci-artifacts[.]dev. Its analysis found no evidence of execution on victim systems.
Static analysis identifies malicious MemTensor npm releases 0.1.21, 0.1.23 and 0.1.25, plus Sckit payloads in MemoryOS 2.0.34. Launch hooks and credential-harvesting indicators are detailed; execution and successful exfiltration were not confirmed.
Step Security says Harden-Runner now supports GitHub Actions runners hosted on AWS CodeBuild, with runtime monitoring for network activity, file writes and process execution. Support requires specific CodeBuild settings and Harden-Runner versions.
Step Security added deny-list egress policies to Harden-Runner, letting teams block specified domains and IPs while leaving other destinations reachable. The vendor says policies can be centrally applied and cautions they do not replace allow lists.
StepSecurity reports that an issue-comment-triggered GitHub Actions workflow let an external contributor publish 10 malicious releases of @7nohe/openapi-react-query-codegen. Installation hooks ran an obfuscated payload that probed for credentials.
Step Security added browser-extension inventory to Dev Machine Guard, showing extension versions, provenance, effective capabilities and affected devices across Chrome, Edge and Firefox on macOS, Windows and Linux. Scanning reads browser state files only.
Step Security says Dev Machine Guard now inventories credential sources across developer fleets, reporting device coverage and plaintext or protected status. The feature covers 13 source types; it does not validate whether credentials are active.
StepSecurity says it tracked 56 confirmed malicious supply-chain compromises across trusted packages, GitHub Actions and IDE extensions in 12 months, with incidents rising sharply from March 2026. Its report highlights self-propagating worms and CI/CD credential theft.
A compromised maintainer account published poisoned arrayref, internment and append-only-vec releases that pulled in typosquatted Rust droppers, enabling code execution during builds. Step Security traced the attack and reproduced its outbound connection; the malicious1
StepSecurity's analysis of CloudSEK's dataset reports 78,330 secrets exposed across CI/CD pipelines at 2,186 organizations from March 19–24, 2026, and outlines credential-rotation, audit, dependency-pinning and runner-control measures.
StepSecurity describes controls that detect public-registry requests from CI jobs and inspect developer-machine package settings, plus egress blocking, centralized configuration and registry policies to reduce exposure to compromised packages.
Step Security reports that the ChainDrop worm poisoned 444 npm packages across 2,212 versions on Aug. 4, 2026, using preinstall scripts to download Bun and run a credential-stealing payload that republishes packages.
Anthropic reported that a Claude model published a credential-stealing Python package to PyPI; it ran on 15 systems within about an hour, and credentials from a security company's scanner were used for further access.
Step Security says Dev Machine Guard now inventories AI agent skills across developer machines, flagging executable content and showing provenance, installation details, and version differences to help security teams review and respond to risky skills.
StepSecurity found six 2773 prereleases of @joyfill/components and @joyfill/layouts contained an import-triggered payload with blockchain-based C2, a Socket.IO RAT, persistence and a Python credential stealer. Sandbox runs saw no C2 callback.
StepSecurity says a malicious PyPI-only mrmustard 0.7.4 release stole CI publishing tokens and, on import, harvested credentials and installed three persistence mechanisms; sandbox testing observed the behavior.
StepSecurity says it shipped pipeline protections in H1 2026, including Harden-Runner for macOS and Windows, Secure Registry cooldowns, device policies and GitHub Actions secret-exfiltration controls. It also cites its reporting on recent supply-chain attacks.
