All 66 cases

Documented case · 2026

Paul Kuo, Taiwan

Suspended Apr 29 and Jun 22, 2026

UnresolvedAbout 2 weeks of disruption from the first suspensionCode access lost / 404CI/CD and deploys

Verified primary source

The account holder's own post, discussion or thread was opened and read in full for this entry.

Suspended twice in eight weeks. The first time, on April 29, 2026, his pushes began failing and the account was "suspended for a Terms of Service violation," with no warning and no specifics. The broken chain was his whole publishing life: writing, to a GitHub repo, to Actions, to Cloudflare Pages. Local drafts survived; publishing did not. He spent about two weeks rebuilding with mirrors and a decoupled deploy, and was publishing again by May 12. The second suspension, on June 22, produced the same 403s, and git push, dashboard sync, and every GitHub-identity operation (CLI and API) failed, but his site and deploys never went down. His second post names the blind spot the rebuild exposed: two AI assistants that looked like independent paths both depended on the same GitHub identity, so both failed together, a same-source single point of failure.

The record

Suspended or flagged
April 29, 2026, and again June 22, 2026
Posted
His blog, May 15, 2026 and June 22, 2026
Ticket / discussion
None published
Suspected trigger
Not stated either time: no advance warning, no specific explanation, and no appeal response as of the first post
Human reply from GitHub
No appeal response as of May 15, 2026, per his first post
How it ended
Not stated as a GitHub restoration in either post. He rebuilt around the first suspension in about two weeks (Codeberg and GitLab mirrors, object storage, a decoupled deploy), and when the second suspension came, the new setup absorbed it: the website never went down and the newsletter went out on schedule
Duration
Under 1 month

Suspected triggers are the account holder’s inference, never GitHub-confirmed, unless the record above says GitHub stated or confirmed it. A suspension does not by itself prove GitHub made a mistake.

Sources

First-hand posts by Paul Kuo, paulkuo.tw, May 15, 2026 and June 22, 2026; both opened and checked November 17, 2026.

His posts document the workaround, not a GitHub decision; whether the account itself was ever restored is not stated.

This entry is one of 66 in What Kind of Life Does a Flag Preserve?, a documented account of GitHub suspending a fourteen-year-old developer account and the same pattern repeating across the industry. If the same thing has happened to you, the article’s eight-step playbook is built from what actually moved these cases.