Nixpkgs core team disbands, citing steering committee friction

The two people responsible for governing Nixpkgs quit last night. In a post published to the NixOS Discourse at 21:39 UTC on August 7, the Nixpkgs core team announced it "has unfortunately decided to disband," roughly ten months after the project's Steering Committee created it. The matters in its jurisdiction, the post says, are "left with no direct owner at present."

That is a bigger deal than a two-person resignation normally is. Nixpkgs is the package set behind Nix and NixOS, one of the largest software collections in open source, and as of this morning its GitHub repository carries 12,333 open pull requests and roughly 9,300 open issues. The layer meant to break deadlocks above that queue is now empty, and it emptied itself with a Steering Committee election due imminently.

RelatedPython 3.15 Betas Land as CPython Puts Its JIT on a Clock

• Who left: the two remaining core team members, @alyssais and @emilazy. The team launched in September 2025 with four.
• Why: the role turned out to be incompatible with writing code, recruitment produced a single active applicant, and the team says the Steering Committee micromanaged it while communicating poorly.
• What it leaves behind: 19 newly onboarded committers, a reformed committer delegation process, an extended merge bot, and the project's first automation and AI policy, none of which now has a maintaining body.
• What happens next: the Steering Committee is the only remaining backstop for Nixpkgs governance until it either re-delegates the work or absorbs it.

What did the core team actually build in ten months?

The Steering Committee stood the team up on September 24, 2025, handing it four constitutional powers wherever they touched Nixpkgs: project direction, decision-making, coordination with the NixOS Foundation Board, and the creation and management of teams. Four people signed on. The pitch was explicitly lightweight. Members said at the time that they intended to keep doing ground-level work and did not want governance to become a full-time job.

The output list is not small. Over ten months the team reformed how commit access is delegated and onboarded 19 new committers, extended the merge bot so maintainers could land changes without waiting on someone holding the right bit, re-established a working relationship with GitHub and secured a sponsored Enterprise Cloud upgrade for the organization, helped triage the GHSA-67f2-674w-6g63 advisory and the GitHub security exposure it revealed, and shipped an initial automation and AI policy. That last item is the one most likely to be overlooked in the fallout.

Why did they quit?

Not over a blow-up. The post is explicit that the decision "is not the result of any one incident, but the culmination of long-running patterns," and that the two remaining members reached it two weeks before publishing.

Three threads run through the explanation. The first is workload: the job "has sadly not turned out to be the lightweight role compatible with active technical contribution that we had originally hoped it would be." The second is recruitment. The team put out a call for new members in May and got exactly one person actively applying, with what it describes as a mixed response to direct outreach. Attrition had already taken the team from four members to two. That arithmetic does not support a functioning body.

The third thread is the sharp one. The team says the Steering Committee "lacks a native instinct for the delegation envisioned by the constitution, while also not being sufficiently engaged and cohesive to handle individual decisions at those levels itself." In practice, they write, that meant matters arriving with the desired outcome already attached, the Committee taking ownership of issues sitting squarely inside delegated areas, no clarity about whether a given member was speaking personally or for the body, and delayed responses to concerns raised repeatedly. The named casualties: coordination on Google Summer of Code, grant initiatives, and AI policy, plus slow progress on moderation and GitHub organization owner reform.

Why does an open-source AI policy need an owner?

Buried in the resignation is the team's proudest result, and it is the one with the shortest shelf life without them. Nixpkgs adopted an initial automation and AI policy this year, covering how machine-generated contributions are handled in a repository that takes patches from thousands of people. The team points to the strong approval it drew "from people with highly divergent views" as proof that even intractable topics can move.

Policies like that are not self-executing. Someone has to answer the edge cases: whether a specific tool counts as automation, what disclosure a contributor owes, what happens when a maintainer refuses a machine-written patch on principle. Every large repository is being forced to write these rules right now, and most are learning what Nixpkgs just demonstrated, that drafting the policy is the easy half. With the core team gone, questions about the AI policy escalate to the same Steering Committee the team singled out for slow coordination on AI policy in the first place.

Who is affected, and what breaks first?

Not the average NixOS user, and not this week. Package updates keep flowing, the merge bot keeps merging, and the maintainer teams doing the daily work are untouched. Governance failures in open source are slow-moving. The damage shows up as unresolved disputes, not as a broken channel.

RelatedVue 3.6's Vapor Mode Takes Direct Aim at Svelte and Solid

The exposed group is contributors with something contested in flight. Anyone who had a conflict escalated to the core team, or who was waiting on a call about a large cross-cutting change, now has no venue below the Steering Committee. The committer pipeline is the other pressure point. Nineteen people were granted commit access through a process the departing members designed, and that process has nobody assigned to run the next round.

Both members say they plan to reduce their Nixpkgs involvement, and neither intends to run for the Steering Committee. The institutional memory leaves with them.

1. 2025-09-24Steering Committee establishes the Nixpkgs core team Four members, four delegated constitutional powers
2. 2025 to 2026Committer delegation reformed, 19 new committers onboarded Merge bot extended, sponsored GitHub Enterprise Cloud upgrade secured
3. 2026-05-11Initial automation and AI policy goes out for feedback The same update calls for new team members
4. late July 2026The two remaining members decide to step down Described as the culmination of long-running patterns
5. 2026-08-07The team announces it has disbanded 21:39 UTC, NixOS Discourse
6. nextSteering Committee election Decides whether Nixpkgs governance gets re-delegated

Our take

The most quotable line in the post is also the most uncomfortable one for the community it addresses. The team argues that a combative, zero-sum posture toward governance, understandable given the project's history, "only rewards those who don't care to listen to the community," burns out the people who do engage in good faith, and "risks locking in the historical status quo of decision-making by deadlock and attrition." NixOS built the Steering Committee and its constitution as the answer to the governance crisis that tore through the project in 2024. The first team handed real power under that constitution has now walked out saying the delegation never actually happened.

Cachix founder Domen Kožar, replying within hours, put the stakes plainly: "This would be the time where nixpkgs should glow and grow."

The coming election is the test. Treat this as a staffing problem and the vacancy stays a vacancy. Treat delegation as something that has to be practised rather than written down, and the structure the core team built is still sitting there, documented, with 19 committers who came through it.

Original analysis by GenZTech. Reporting informed by the NixOS Discourse announcement.