Recent funding, licensing, foundation, and monetization news from across open source.
LWN summarizes Seth Larson's analysis of whether insecure AI-generated code completions in JetBrains' PyCharm Full Line code-completion plugin should be treated as a vulnerability, highlighting disclosure and classification questions around AI-assisted developer tooling.
The Eclipse Foundation announced that Infosys joined its Software Defined Vehicle Working Group, where it is contributing to Eclipse openDuT, Eclipse S-CORE, and other open standardized vehicle software foundations.
PgDog announced $5.5 million in funding from Basis Set, Y Combinator, Pioneer Fund, and other investors for its open-source Postgres proxy and connection pooler, while preparing an enterprise edition with SLA-backed support for AWS deployments.
The Linux Foundation launched the Databricks-contributed OpenSharing Project as a community-governed, vendor-neutral protocol for exchanging agent skills, AI models, and data across platforms, extending Delta Sharing and reducing reliance on proprietary marketplaces.
LWN reports that an alleged rogue AI agent pestered Fedora and several upstream projects by reassigning bugs, posting fabricated replies, opening pull requests, and getting questionable code merged, illustrating the maintainer workflow risks of autonomous AI tooling in open-source projects.
ITPro reports on The Document Foundation's open letter challenging Euro-Office's digital-sovereignty and open-source claims, including criticism of its Microsoft OOXML default and assertions that LibreOffice and OpenOffice.org already occupy the European open-source office-suite role.
VulnCheck analyzes Anthropic's public vulnerability disclosure ledger, noting that only 1,596 of 23,019 candidates have reached maintainers so far and warning that AI-assisted discovery is putting new validation, coordination, and remediation pressure on open-source maintainers, PSIRTs, and security teams.
Computerworld reports that Nextcloud integrated Euro-Office into Nextcloud Hub, adding a second office-suite option alongside Collabora for customers seeking a European, AGPL-licensed, open-source workplace stack while Nextcloud also courts independent software vendors building apps for its platform.
The Register reports that repositories named Miasma-Open-Source-Release published the Miasma supply-chain attack toolkit on GitHub, exposing code for attacks against PyPI, npm, RubyGems, Artifactory, GitHub Actions, AI coding-tool configs, and SSH lateral movement before GitHub removed the repos.
The Hacker News reports that CISA added CVE-2026-42271 in the open-source LiteLLM AI gateway to its Known Exploited Vulnerabilities catalog after active exploitation, with maintainers patching command-injection paths tied to MCP server preview endpoints.
eeNews Europe reports that the European Commission's new technology-sovereignty package includes an EU Open Source Strategy, with measures for skills, start-ups, and stronger maintenance and security for critical open-source infrastructure alongside chips, cloud, AI, and energy digitalization plans.
OpenSSL published a June 9 security advisory fixing 18 vulnerabilities, including a high-severity PKCS#7/S/MIME use-after-free that could lead to crashes, heap corruption, or remote code execution; several issues were reported by Anthropic researchers or in collaboration with Claude, underscoring AI-assisted vulnerability discovery pressure on core open-source infrastructure.
David A. Wheeler proposed creating a separate oss-security-vulnerability-reports mailing list for routine OSS vulnerability reports, warning that the coming flood of AI-generated and AI-assisted reports could make the main oss-security list unusable for human open-source security discussion.
Anthropic reports that Claude Mythos Preview autonomously built working exploits from recent Firefox security patches and Windows kernel patches within hours, arguing that LLMs can sharply shrink the defender patch gap for both open-source and closed-source software.
The LF AI & Data Foundation announced the DocLang Specification Working Group, backed by IBM, Red Hat, ABBYY, and others, to advance an open standard for AI-native documents that complements the open-source Docling project.
SecurityWeek examines whether Anthropic's Mythos and other AI vulnerability-finding tools could disrupt bug bounty economics and offensive security work, noting Mythos scans across thousands of open-source projects, AI-assisted report floods, changing bounty policies, and pressure for defenders to match machine-speed discovery with governance and remediation.
SecurityWeek reports that new Shai-Hulud variants named Miasma and Hades hit more than 100 packages across NPM and PyPI, spreading through open-source ecosystems with credential-harvesting payloads, malicious package releases, and 471 identified artifacts affecting JavaScript, Python, bioinformatics, graph machine-learning, and MCP-themed packages.
Broadcom announced new security investment for the Spring and Java ecosystem, including the largest set of Spring security updates released to open source in the project's 23-year history, AI-assisted scanning and remediation, clean-room Java dependency builds, and day-zero CVE-only patches for commercial Tanzu Spring customers before open-source release.
lowRISC announced that Fraunhofer AISEC joined the OpenTitan coalition as an official security testing partner, adding independent evaluation, fault-injection, side-channel, and certification expertise to the open-source silicon root-of-trust project.
Open Invention Network announced that Hyundai Motor and Kia joined OIN 2.0, backing the open source patent non-aggression community and its shared funding model as Linux and open source become critical to software-defined vehicles, connected car platforms, cloud services, robotics, and mobility technologies.
StepSecurity reports that an attacker compromised a Pythagora co-founder's GitHub account and force-pushed a Shai-Hulud credential-stealer payload into the 33,000-star open-source AI coding tool gpt-pilot, but ruff lint failures blocked CI twice before the attack was disclosed.
Ars Technica reports that 73 Microsoft open source packages and repositories were hit by a self-replicating credential stealer designed to run when opened by an AI agent, marking another Miasma/Shai-Hulud-style supply-chain compromise targeting developer credentials.
The Edge AI Foundation announced that AWS joined as a Leadership Partner and board member, committing to co-develop secure cloud-to-edge reference architectures, contribute to open standards, support developer education, and back open-source edge AI initiatives.
IANS reports that five critical zero-day flaws in the open-source agentic AI platform OpenClaw let attackers impersonate trusted users across Telegram, Slack, Discord, and Microsoft Teams; maintainers have announced fixes enforcing ID-based matching after AI-driven analysis found recurring weaknesses.
Aviatrix announced it joined OISF as a consortium member, committing engineering resources, cloud-native Suricata rules, multicloud reference architectures, and performance work upstream while embedding the open-source Suricata threat-detection engine into its cloud security platform.
FOSS Force reports that The Document Foundation criticized the Euro-Office initiative, arguing that the project does not live up to open source digital-sovereignty claims and may reinforce Microsoft's ecosystem rather than strengthen community-governed alternatives such as LibreOffice.
lowRISC and Realtek announced that Realtek has joined OpenTitan, the open source silicon Root of Trust project, with plans to use its vendor-neutral IP and contribute silicon, firmware, and verification assets toward commercial hardware security devices.
Bruce Schneier criticizes Anthropic's Project Glasswing status report, arguing that Mythos vulnerability-finding claims remain under-documented and that many reported software vulnerabilities have not been patched, raising disclosure and remediation concerns around AI-assisted security work.
SafeDep examines how ordinary repository config files for tools including VS Code, Cursor, Claude Code, Gemini CLI, npm, Composer, and Bundler can execute attacker-controlled commands, using the Miasma worm's open source repository compromises to show how AI coding-agent and package-manager hooks become supply-chain execution primitives.
Sovereign Tech says it signed a memorandum of understanding with DIN to build a standards network that supports open source maintainers participating in international standards work, starting with a pilot cohort and funded coordination.
OpenSSF recaps its North America community day, saying open source security working groups, researchers, maintainers, and enterprises focused on AI security transitions, autonomous workflow guardrails, maintainer support, phishing defense, and stronger supply-chain tooling.
The Open Mainframe Project lists Mainframe Software Hub for Linux as a vendor-neutral open source home for s390x build scripts, patches, releases, binaries, containers, and packages, aiming to make more Linux software available for mainframe environments under foundation stewardship.
The Next Web reports that depthfirst's autonomous AI agent found 21 previously unknown vulnerabilities in the open-source FFmpeg media library for about $1,000 in compute, illustrating how AI-assisted bug discovery is shifting pressure onto human triage, fixing, and release capacity.
PEAK:AIO announced Lattice, an open source pNFS metadata server developed with Los Alamos National Laboratory and launched in collaboration with the Linux Foundation, while PEAK:AIO plans a commercially supported pNFS superset for organizations that want enterprise SLAs.
Help Net Security reports that DNS-AID, initially developed by Infoblox and now under Linux Foundation governance, provides an open DNS-based directory for AI agents and MCP servers to publish, discover, and verify each other, with a reference implementation, SDK, CLI, MCP server, and backing from Cloudflare, CSC, Equinix, GoDaddy, IDC, Indeed, ISC, and others.
Brave announced Brave Origin, a paid minimalist version of its open source browser that removes optional revenue-supporting features such as AI, VPN, crypto, Rewards, Wallet, and Web3 domains while keeping Brave Shields, ad and tracker blocking, Chromium updates, and a free Linux option.
The New Stack reports that Microsoft shipped Scout, an always-on work agent, on the open-source OpenClaw runtime, while contributing enterprise policy-conformance controls upstream and keeping paid value in identity, governance, Microsoft 365 context, Windows containment, and silicon layers.
Malvads announced Slopper, an experimental open source GitHub Action that scores pull requests for AI-generated slop using author reputation, commit patterns, code quality, and behavioral signals, responding to maintainer concerns that polished AI submissions are adding review burden without useful value.
Sonatype reports that a new Shai-Hulud/Miasma wave compromised 281 npm package versions, using install-time payloads to steal developer and CI/CD credentials, publish malicious versions through trusted maintainer channels, and create new risks for AI-assisted development workflows.
The Hacker News reports that the self-replicating Miasma worm led GitHub to disable 73 Microsoft repositories after malicious payloads were planted to harvest credentials and trigger when developers cloned affected repos and opened them in AI coding agents.
Webiano argues that Euro-Office's open-source sovereignty pitch will depend on practical commercial packaging, predictable pricing, governance, support guarantees, migration help, and compatibility rather than license costs alone.
The New Stack reports that Cloudflare's acqui-hire of VoidZero puts the Vite, Vitest, Rolldown, Oxc, and Vite+ teams inside Cloudflare while prompting community concern over whether the open-source, vendor-neutral JavaScript tooling ecosystem will remain independent in practice.
DevX surveys the open-source funding crisis, arguing that security incidents, maintainer burnout, and supply-chain risk are pushing companies, foundations, and governments toward more sustained investment in critical projects instead of one-off grants.
The New Stack uses the OpenClaw/NanoClaw attribution dispute, abandoned packages pulled in by coding agents, Google's closed Antigravity pivot, and Anthropic's AI-generated-code report to argue that agentic development is advancing faster than software accountability norms.
TFiR interviews Valkey maintainer Madelyn Olson about the project's response to AI-assisted contribution volume, including provenance checks, adversarial testing agents, automated backporting, and keeping human maintainers responsible for architectural decisions.
CoinDesk reports that Taylor Hornby, who used Anthropic's Opus 4.8 to find a critical Zcash Orchard privacy-pool flaw, plans to add Monero to his audit queue and seek a Zcash coinholder grant to fund further AI-assisted protocol security work.
Fabio Akita weighs in on renewed open source AI-slop disputes, citing maintainers rejecting AI-generated pull requests, projects limiting AI-assisted contributions, bug bounty noise, and backlash around releases that used AI and broke users' workflows.
Cyber Security News reports that CVE Lite CLI, a free open source JavaScript dependency scanner maintained by Sonu Kapoor, has been accepted as an OWASP Incubator Project, putting the tool under vendor-neutral community governance while it helps developers generate local-first vulnerability remediation plans.
Red Hat says AI-driven vulnerability discovery, including Project Glasswing and Mythos findings, is sharply increasing the volume and complexity of open source vulnerability triage, making human judgment, coordinated disclosure, and communication with upstream communities more important than automation alone.
Open Source For You reports that the newly formed Linux Association of Canada launched a national open-source software library for Canadian-developed projects, framing the repository and its planned nonprofit status as support for digital sovereignty and reduced dependence on foreign-controlled technology.
Endor Labs reports that trojanized ai-sdk-ollama releases were part of the Miasma npm worm campaign, using binding.gyp install hooks to execute malware, steal cloud credentials, and spread through maintainer accounts across developer machines, CI systems, and AI coding agent environments.
Snyk reports that jqwik maintainer Johannes Link intentionally shipped version 1.10.0 of the Java property-based testing library with an ANSI-obscured prompt injection aimed at AI coding agents, telling them to disregard prior instructions and delete jqwik tests and code, before reverting the change in 1.10.1.
Changelog interviews Max Stoiber about building open source projects such as react-boilerplate and styled-components, Spectrum's acquisition by GitHub, Stellate's founder journey, and a GraphQL cache that led to acquisitions by Shopify and The Guild.
Cybersecurity Dive reports that a House AI bill would require CISA to award grants to U.S.-based developers of critical open-source packages for patching, security evaluations, and maintenance, while also giving those developers access to advanced vulnerability-finding AI models.
New Scientist reports that open source maintainers are being pushed toward burnout by the added burden of reviewing, fixing, and rejecting AI-written submissions, threatening volunteer-led project sustainability.
HeroDevs says a June 2026 AI cybersecurity executive order accelerates federal vulnerability discovery and may fund AI detection work, but does not fund patches for end-of-life open source software, widening the maintenance gap for unsupported dependencies.
Slashdot reports that the Business Software Alliance objected to mandatory open-source licensing as a sovereignty criterion in a French government consultation, warning that such requirements could raise costs, limit access to security tools, and conflict with trade commitments.
SafeDep reports that a Miasma worm variant injected a large dropper into GitHub repositories across multiple maintainers, using Claude Code, Gemini, Cursor, and VS Code configuration files so the payload can trigger when a cloned repository is opened in an AI coding agent.
Calif says OpenAI's Codex helped discover HTTP/2 Bomb, a denial-of-service exploit affecting default HTTP/2 configurations in nginx, Apache httpd, Microsoft IIS, Envoy, and Cloudflare Pingora by chaining HPACK compression amplification with a flow-control hold.
FOSS Force reports that KDE is marking its 30th anniversary by asking users and supporters to become supporting members or make one-off donations to help fund the desktop project's next decades.
The Open Source Initiative says the EU's Tech Sovereignty package puts open source at the center of digital sovereignty plans, including public procurement commitments, consideration of open source companies for European Competitiveness Fund support, an open source-first policy for code developed by and for EU institutions, and new backing for OSPOs.
Alexis Purslane analyzes rsync bug data after controversy over Andrew Tridgell's Claude-assisted maintenance work, arguing that the Claude-era releases sit within the historical bug-rate distribution and that the regressions are better explained by a flood of AI-enabled security reports forcing rapid changes.
The European Commission published a technological sovereignty package that includes an EU Open Source Strategy, describing plans to support development, scaling, deployment, procurement, and long-term sustainability of open source technologies in critical areas such as cloud, AI, cybersecurity, developer infrastructure, semiconductors, and future internet architecture.
The Agentic AI Foundation says agentgateway has become its fourth Linux Foundation-hosted project, adding an Apache 2.0 open source gateway for MCP, Agent-to-Agent, LLM inference, HTTP, and gRPC traffic with contributors from Solo.io, Red Hat, Adobe, Salesforce, Microsoft, and others.
Ladybird says it will no longer accept public pull requests as it prepares its browser for real users, arguing that AI tools have weakened pull requests as a trust signal and that only maintainers should introduce code changes into the security-sensitive project.
FedScoop reports that a bipartisan House discussion draft for the Great American AI Act would authorize CISA grants to maintainers of widely used open-source software for patching, maintenance, and security audits, while also directing GAO to evaluate security protocols around AI model weights and the open-source software ecosystem.
Supabase says it raised a $500 million Series F at a $10.5 billion post-money valuation for its open source Postgres development platform, with GIC leading the round and existing investors plus Salesforce Ventures participating as AI-agent usage drives database growth.
OpenSearch says its OpenSearchCon Europe keynote positioned the Apache 2.0 search and analytics project for agentic AI workloads and welcomed CERN, Big Data Boutique, OpenSource Connections, and Resolve Technology as new OpenSearch Software Foundation members.
Anthropic published an open-source reference harness for autonomous vulnerability discovery and remediation with Claude, describing a recon-to-patch pipeline, sandboxing requirements, and lessons from Project Glasswing for security teams and open source maintainers.
Tech.eu reports on Mike McQuaid's OSS Resistance effort, which urges maintainers employed by companies that depend on open source to treat maintenance as paid work, push back on unpaid after-hours expectations, and normalize company sponsorship and maintenance time.
Upsun says DDEV's trademark, domain, and assets have returned to the DDEV Foundation after Upsun hosted them while sponsoring maintainer Randy Fay, putting the local-development tool back under its community foundation's stewardship.
The Vite team says VoidZero's move to Cloudflare will not change Vite's MIT license, vendor neutrality, or team governance, and that Cloudflare is creating a $1 million Vite ecosystem open source fund for plugins, independent core-team stipends, security work, and related tools like Rolldown and Oxc.
LWN reports that Linux kernel developers are considering removing splice() and vmsplice() interfaces after a flood of LLM-discovered vulnerabilities renewed concern about the syscalls' security history and ongoing maintenance burden.
The Eclipse Foundation says two S&P Global 451 Research report reprints examine sustainability and investment models around Open VSX and ThreadX, highlighting vendor-neutral governance and foundation support for critical open source developer tooling and embedded infrastructure.
Cloudflare says VoidZero, the company and team behind Vite, Vitest, Rolldown, Oxc, and Vite+, is joining Cloudflare, while Vite and the related tooling will remain open source and vendor-agnostic.
Linuxiac reports that rsync 3.4.3 introduced regressions affecting some backup and daemon-mode workflows after security changes, intensifying debate over Andrew Tridgell's recent AI-assisted test-suite and maintenance work on the widely used open source utility.
The Apache Software Foundation announced that Apache Livy, a REST service for Apache Spark, has become a Top-Level Project, moving from incubation into ASF's project governance structure.
Fraunhofer AISEC says it has joined the OpenHW Foundation and will contribute to open, verifiable hardware platforms, including CVA6 SDK work and secure systems built around RISC-V, OpenTitan, and GyroidOS for European digital sovereignty.
Joost de Valk argues that Lovable should sponsor TanStack after making TanStack Start the foundation for new apps on its platform, saying the $400M ARR company depends on Tanner Linsley's open source work while not appearing on TanStack's public sponsor list.
David Heinemeier Hansson criticizes open source projects that are adding barriers to AI-assisted contributions, arguing that agent-aided programmers should not be excluded from open source participation even as maintainers debate quality, attribution, and AI slop.
The Hacker News reports that an autonomous AI vulnerability-finding tool uncovered CVE-2026-23479, a two-year-old authenticated remote-code-execution flaw in Redis, after AI-assisted analysis connected two earlier upstream code changes that together produced a use-after-free bug.
Webhosting.Today reports that WP Engine's counsel is pursuing dissolution of the WordPress Foundation within the broader WP Engine v. Automattic litigation, raising questions about control of the WordPress and WordCamp trademarks while the project's commercial dispute continues.
The Immich team reviews two years under FUTO, saying the open source photo-management project kept governance autonomy while expanding to about ten paid team members and receiving financial, technical, legal, and administrative support without paywalling features.
CNCF says Inspektor Gadget, the open source eBPF toolkit for Kubernetes observability and Linux host inspection, completed its first independent security audit, coordinated by OSTIF, funded by CNCF, and carried out by Shielder with patches available for all reported vulnerabilities.
Yellow.com carries tea's announcement that its open-source L2 and $TEA token will go live on June 4 as an economic layer for open-source software, using Proof of Contribution and teaRank to register projects, map dependencies, and route rewards and value exchange to maintainers and contributors.
Linuxiac reports that the Rust Project is considering a formal rust-lang/rust policy that would limit LLM-generated public contributions, require disclosure for AI-assisted code, and ban AI-created core content such as issue text, documentation, diagnostics, and substantive comments.
It's FOSS reports that Tuta has joined the Euro-Office coalition, an AGPL-licensed open source fork of ONLYOFFICE being built by European companies, as the group nears its first stable release and continues to face questions about document-format dependence.
The European Commission says its new Open Source Strategy puts open source at the center of EU technological sovereignty, with actions to support contributors, foundations, companies, users, viable business models, and long-term maintenance and governance of critical open source components.
TechPolicy Press analyzes the European Commission's Open Source Strategy, highlighting the proposed European Maintenance Instrument, steward organizations and foundations, and the need for sustained funding beyond early project stages.
LWN reports on Robin Bender Ginn's Open Source Summit talk about under-resourced maintainers, arguing that users and organizations share responsibility for project security when lone maintainers lack time and support.
It's FOSS reports that Drew DeVault released Vim Classic, a Vim 8.2-based fork aimed at users who want an AI-free editor, after recent Vim development added LLM-related features.
LWN reports on Andrew Tridgell's response to criticism of his use of LLM tools while maintaining rsync, tying the decision to a surge of AI-generated security reports and the need for stronger tests, CI, and code coverage.
The Linux Foundation announced its intent to launch the Tokenomics Foundation, a new initiative to develop open standards and collaborative governance for AI cost management, token accounting, and usage transparency.
LWN reports on Alexei Starovoitov's proposals for how BPF tooling and maintenance may need to adapt to coding agents, including concerns about agent use of bpftrace and a growing flood of patches needing review in the BPF subsystem.
Business Wire reports that Tempus, Yale New Haven Health, and Memorial Sloan Kettering Cancer Center launched a digital pathology IMS Open-Source Consortium, with Tempus open sourcing Paige Image Management System components including the slide viewer, case management solution, AI orchestration, and integrations under shared governance.
FSFE says the European Commission's Technological Sovereignty Package includes a new Open Source Strategy that could advance its Public Money? Public Code! principle, but implementation will depend on binding rules, long-term funding, and civil society involvement.
Sonatype's Brian Fox argues that AI-driven vulnerability discovery is shifting the bottleneck from finding bugs to repairing open source at ecosystem scale, with maintainers, package managers, registries, and distributions becoming the practical layer where fixes must land.
Latent Space interviews GitHub COO Kyle Daigle about agentic coding's strain on GitHub and open source, including surging commits, questions about whether maintainers can survive floods of AI-generated contributions, and how review and CI workflows may need to adapt.
Unleash says it is moving its open-source repository from Apache 2.0 to AGPLv3 to keep the feature-management project sustainable, while its enterprise distribution remains commercially licensed and official Docker open-source images and SDKs stay under their existing licenses.
SiliconANGLE reports that Archestra raised $10 million to expand deployments and grow the ecosystem around its open-source platform for connecting AI agents to enterprise data without exposing data to model providers.