Source Verification: Snapshots read from disk (gunzip, sha256 of each decompressed file matched manifest; suspicious_patterns null for all 8). source-0 LWN (Brockmeier, 6 Oct 2026): confirms every Gentoo claim in the body (last rites 24 Sep, period to 24 Oct, Zilka bug 9 Sep, James/Jolly quotes verbatim, three-people estimate, release cadence, Fedora 154.0.8037.92 vs upstream .97 of 1 Oct, Debian security repo same, Flathub Flatpak third-party, Keeping up with Chromium is not an easy task, bug 982204). Its page carries Gentoo developers' own comments (sam_c, juippis), so those are not independent, as the article says. source-1 Linuxiac (Borisov, 18 Oct): removal scheduled 24 Oct unless a volunteer, package already masked, Chrome stays as precompiled binary, 'more than six hours of personal time each week' for the current maintainer, at least three developers: all confirmed; it summarises the mailing-list announcement and bug tracker (restates). source-2 GitHub commit 00ba5be: snapshot text confirms title 'www-client/chromium: treeclean', Bug 982204, Signed-off-by Matt Jolly, 17 files, 14,945 deletions, but carries no date; I read the date from the GitHub API (gh api repos/gentoo/gentoo/commits/00ba5be...): author date 2026-11-05T07:27:13Z, committer date 2026-11-05T08:37:40Z, so 2026-11-05 UTC holds on either basis; 24 Oct to 5 Nov is 12 days. The commit removed www-client/chromium and www-client/chromium-common; www-client/google-chrome (157.0.8103.x ebuilds) remains in the tree and no chromium-bin exists (404), consistent with LWN's note that Gentoo offers no binary Chromium. 'Removed' vs 'treecleaned': the article says removal and names the commit title 'treeclean', which is the same act. source-3 Phoronix via Archive.org capture of 2026-10-23 (manifest archive_fallback true; byline 'Written by Michael Larabel ... on 1 August 2026 at 07:47 AM EDT', matches the article's US Eastern attribution): Linderud resigning after a decade; 'it's a good time to let go'; 'among the Arch Linux packages now in need of a new maintainer include the likes of mkinitcpio, Bolt, nvme-cli, Go-Tools, Pacman, arch-install-scripts, archlinux-keyring, archlinux-repro, and others'; 'also the sole maintainer of WPA_Supplicant, fsverity-utils, and other packages'; 'many other packages'. 'Left packages needing owners' is Phoronix's own framing of his resignation email, not a Herald inference; the article does not repeat the earlier Herald headline error about Pacman (corrections record 2026-08/04-prominent-arch...): Pacman appears only in the list Phoronix gives. source-4 Register (Proven, 13 Nov 15:28 UTC): 'isn't quitting', four maintainerships quote from his CV, the LKML quote verbatim ('Given the current throughput ... except in those areas I maintain.'), 'I am retaining my maintainerships and have no plans on giving those up, I just want my inbox to be less horrendous', 'once a week' to 'daily' to 'half of what I see in review', 'across the board', 'just the day after', Kroah-Hartman 'All of the maintainers are burnt out, and we have been for ten years.' and 'I want to go down to two Long Term Support releases - four, maximum. Six is too many.' (the source of the Six-is-too-many quote is this Register article, at the Open Source Summit 'last month'), all confirmed verbatim. I could not read the LKML post itself (lore.kernel.org serves a bot-check page); the article attributes the words to The Register's quotation of it, which is accurate. source-5 Register (9 Nov): Torvalds 7.4-rc2 post, 'Alexei claims that on the bpf side at least the fixes are now coming in faster than the new reports', 'I think most other areas haven't reached that point yet' confirmed; it mentions Anthropic and Mythos once as background and the article does not use that. source-6 kernel.org releases page (live page re-fetched 14 Nov, same table): six longterm branches 6.18, 6.12, 6.6, 6.1, 5.15, 5.10, projected EOL Dec 2028/2028/2027/2027/2026/2026, released 2025/2024/2023/2022/2021/2020; derived lifetimes (about 3 to about 6 years) recomputed and correct; 'falls to four' if Dec 2026 dates hold is correct. source-7 FOSS Force (Hall, 2 March 2026): 6.12 Dec 2026 to Dec 2028 and 6.18 Dec 2027 to Dec 2028, quote 'lots of discussions with different companies and groups and the other stable kernel maintainer' confirmed; restates the announcement and adds a table. HERALD OWN CHECK (Arch package search): archlinux.org is not on the allowlist and is not linked in the body or sources. I re-ran https://archlinux.org/packages/search/json/?name=<pkg> for the ten packages at 2026-11-14 17:12 UTC and reproduced the article's figures exactly: mkinitcpio [lfleischer, grazzolini], bolt [jlichtblau, jelle], nvme-cli [gromit, Antiz], archlinux-keyring [eworm, dvzrv], archlinux-repro [kpcyrd, Antiz] (two each); go-tools [grawlinson], pacman [anthraxx], arch-install-scripts [dvzrv] (one each); wpa_supplicant [heftig] (one); fsverity-utils [] (none). 9 of 10 list a maintainer, none lists Linderud. The article labels the figure the Herald's own, states the date/time and method, and states the limits (current maintainers only, official repositories not AUR, 'many other packages' not checked). Allowlist decision: not added. No cited source rests on archlinux.org and no URL from it is in article.sources, so there is no allowlist warning to resolve and no claim depends on it; adding an unreferenced host would only widen the list. If a later article cites it as a source, add it then with evidence. Independence (policy counts by outlet): outlets independent of the subjects = LWN, The Register (two articles = one outlet), Linuxiac, Phoronix, FOSS Force = 5; original reporting = LWN and The Register; Linuxiac, Phoronix and FOSS Force restate announcements and the article says so; the GitHub commit, kernel.org and the Herald's own Arch query are first-party records or own work and are not counted. Per case: Gentoo = LWN (original) + Linuxiac (restates); kernel review = The Register only (single outlet; Torvalds context from a second Register article, same outlet); Arch = Phoronix only (restates the announcement); LTS = The Register (quote) + FOSS Force (restates) + kernel.org (first party). The policy requirement (at least 2 independent, at least 4 sources, counted by outlet) is met overall; two of the four cases rest on one outlet each, which the article's Sourcing paragraph and its per-case wording make visible. Claim-accuracy tests: ran the absence/quantifier grep on title, summary and body; checked 'none of the cited sources says whether any volunteer has come forward / how Arch assigned the packages / whether Stoakes's change will spread' against all 8 snapshots (true); 'neither Register article gives a measured figure for review workload' (true; the 9 Nov article gives Microsoft/Oracle patch counts and a code-size figure, not review workload); 'All eight packages Phoronix lists ... show at least one maintainer' and 'Of the two named as solely his' (true, recomputed); no 'both/every/only' error found. Dates: Phoronix time zone stated (EDT), commit date stated as UTC, Register dates are UTC bylines, FOSS Force 2 March. Named individuals (Stoakes, Linderud, Kroah-Hartman, James, Jolly, Torvalds): every statement is attributed to the speaker in the source, quotes are verbatim, no motive speculation (the Arch case expressly says no workload reason is given). Anthropic disclosure: not required - the article is not about Anthropic or Claude, does not name them in the headline, summary or lead, and does not use the 9 November Register article's single background mention of Anthropic/Mythos.