GitHub Outage Map

Community Discussion

Tips? Frustrations? Share them here. Useful comments include a description of the problem, city and postal code.

Beware of "support numbers" or "recovery" accounts that might be posted below. Make sure to report and downvote those comments. Avoid posting your personal information.

GitHub Issues Reports

Latest outage, problems and issue reports in social media:

• 
  BullBear.News (@bullbear_info) reported 3 minutes ago

  @github Useful for checking PR status without leaving the terminal, but CLI real estate is precious. Claude Code and Cursor already shifted the bar from context viewing to actual execution. Let's see if v1.0.58 handles large issue backlogs without lagging the shell.

• 
  𝗮𝗰𝗶𝗱𝘃𝗲𝗴𝗮𝘀 (@acidvegas) reported 4 minutes ago

  @zombodb @github A platform with almost 200M user's on it had a little bit of down time? HOW DARE THEY. You sound like the people who start crying as soon as Claude drops to 98% reliability.

• 
  luna (@ImLunaHey) reported 5 minutes ago

  we seriously need a github alternative with how often its down.. kinda over actions not working because the site keeps having issues.

• 
  Bart Ilg (@bart_ilg) reported 8 minutes ago

  @rivdb Yep, most people simply don't have a bunch of machines that need identical configurations. Even if they want to copy parts of their setup across multiple machines, they end up being pretty distinct in roles and thus have alot of unique configurations. Its still amazing though, and when you do need to replicate those machines there isn't anything else comparable. I recently used it for a personal project where I wanted to deploy some small servers for family members. Mostly simple functionality like DNS level ad blocking. Once I got the config set up it became super simple to bootstrap a new server and have it pull updates directly from my github repo, its a dream.

• 
  David putra (@davidputra2112) reported 8 minutes ago

  here's the actual problem $WANE is going after. agents holding their own wallets now sign transactions all day without a human checking each one, and when a drainer gets one of them, nothing about that gets passed on. the same poisoned contract can empty a thousand wallets in a row because every agent is meeting the threat for the first time. the fix is a shared registry. the first agent that gets hit records the threat onchain as an "antibody," tagged by address, contract codehash, and call pattern, so redeploying the same scam at a new address doesn't dodge it. every other agent just calls a free read before signing, check() basically, and a match reverts the transaction before anything moves. anyone can report a suspect address for free too, but it only becomes an enforceable antibody once it's staked in $WANE, and anyone can challenge a false one and get the reporter slashed. $WANE sits behind every one of those records as the bond. it's not a governance sticker, it's what makes lying about a threat expensive. there's also a WaneVault, a non-custodial smart wallet that screens every outbound send against the registry before it clears, and the whole thing is live on both Base and Solana with real repos, tests, and docs on GitHub, not just a landing page. reality check though. this token is hours old, liquidity is around $13.6k against a $33k mcap, and volume in 24h is already north of $250k, so this is trading hard, not sitting still. the code looks like actual engineering, but it's one committer with no outside verification, commits went quiet for about a month before this even launched, and there's no third-party audit on any of it. cool infra idea, early and unproven as a token. treat it that way. CA: HesbMP8FaoUvK8uKrtg6Pf4WCiQXgFS8oQJ3tjZpump DYOR.

• 
  Nyk 🌱 (@nykdotdev) reported 9 minutes ago

  I open-sourced `agent-security` today. Coding agents can clone a poisoned starter, read hostile content, leak private context into a public commit, or run a destructive GitHub command. Each action looks ordinary in isolation. The damage appears when trust crosses a boundary without a gate. I wanted a small preflight layer I could inspect myself, run offline, and place before the expensive security tooling. So I built five controls: 1. `scan` checks a repository before publication for credential shapes, private paths, invisible Unicode, and same-file decode-to-exec patterns. 2. `vet` inspects an incoming template, package, plugin, or skill before install or build. It returns ADOPT, REVIEW, or REJECT and never executes the candidate. 3. `scan-content` adds a prompt-injection defense layer for fetched pages, pasted text, and tool or MCP output. It trips on known instruction-override, exfiltration, credential-solicitation, and covert-action shapes. The tripwire is paired with a behavioral contract: treat untrusted content as data, restrict the capabilities available while processing it, and require human approval before acting on instructions discovered inside it. 4. `guard` puts a local brake in front of destructive `gh` operations. Delete, transfer, rename, archive, and visibility changes require confirmation for the exact repository. 5. `harden` audits the control that matters after local guards fail: whether the active GitHub credential and server-side policy can still destroy a repository. Unverifiable capability returns UNKNOWN, never a green check. The project is deliberately honest about what regex and local hooks cannot prove. A clean scan means no known pattern matched. Novel wording, encoding, translation, and instructions split across lines can evade prompt-injection detection. Cross-file code flow still needs CodeQL or another taint-analysis tool. A direct API client can bypass the GitHub guard. That boundary is part of the product. The repository ships with offline fixture suites, a self-scan, ShellCheck, Gitleaks history scanning, secret scanning, push protection, protected `main`, technical blueprints, and field notes explaining the decisions. Public CI is green. MIT licensed. If coding agents touch your repositories, run the sixty-second quickstart and tell me which boundary I should harden next. Repo in the first reply. Bookmark this. Follow @nykdotev for daily builders talk and OSS.

• 
  Scry (@scrygg) reported 10 minutes ago

  HE TAUGHT CLAUDE TO REMEMBER WHY EVERY GITHUB REPO EXISTS Three months into collecting GitHub repos, one builder reopened the same folder and could not remember why half of it was there. The README explained each tool, but none of them explained his relationship to it 1 vault 2 folders 4 portfolio passes a new audit every 12 hours Then the first review exposed the real problem Some repos were duplicates. Others were marked in-use but appeared nowhere in his projects. A few dependencies had gone quiet upstream while his own code still relied on them So Claude was split into two jobs A cheap model reads every README, manifest, commit date and real project reference. Sonnet compares the entire collection, finds overlapping tools and decides what still earns the disk space The payoff only appears when the system reads across everything 30+ repos become one searchable memory Not just what every tool does but why he grabbed it whether he still uses it and what is quietly starting to rot

• 
  Mirza Aamir (@AmirRasheedBaig) reported 13 minutes ago

  @github the wild part: her team basically invented software engineering as a discipline here. she coined the term because nobody took "software" seriously yet. also the code had error-detection that saved the landing when the computer overloaded during descent. legend.

• 
  Tomek | Builds & Learns (@tomek_builds) reported 13 minutes ago

  GitHub code scanning can now flag security issues in languages and frameworks that CodeQL doesn't support. The findings appear directly on pull requests and are clearly labeled as AI-generated. They won't block the merge, but they may catch issues before the code lands. CodeQL finds what its queries know. AI is now looking beyond that coverage.

• 
  Ever4k (@0xEver4k) reported 17 minutes ago

  I WAS CLONING REPOS, GETTING THEM HALF-WORKING, AND THEN FORGETTING WHY I EVEN SAVED THEM. THREE MONTHS LATER — SAME FOLDER, ZERO MEMORY. I had 30+ cloned GitHub repos sitting around. Some I used every week, some I touched once, and some I probably cloned twice under different names. Every time I opened the folder I felt guilty and lost. READMEs didn’t help — they tell you what the author built, not why I grabbed it or whether it’s still useful. One day I saw a simple system: one Obsidian vault + Claude that reads every repo and writes a personal note for each one (why I grabbed it, how I actually use it, status). Then it runs passes across all notes to find duplicates and dead projects. The first time I ran it on my own clones I was shocked. Half of them were duplicates. Several tools I thought I was using hadn’t had an upstream commit in over a year. And one repo I was actually depending on was basically abandoned. Now instead of a messy graveyard of folders I have a clean list: what I’m really using, what’s duplicated, and what’s quietly rotting. All in plain markdown I can open anytime. The value isn’t in any single note. It’s that nothing can quietly duplicate or die without me noticing. Full setup + exact loops and prompts is in the article in my profile. I break down practical AI systems like this every day — follow so you don’t miss the next one.

• 
  Alejandro Wurts (@Alwurts) reported 17 minutes ago

  Damn @github actions are down, I cannot ship!

• 
  Guilherme Politta (@ThePolitta) reported 20 minutes ago

  And @github is down again…

• 
  nik skld (@nikskld) reported 23 minutes ago

  ASKED FOR THE BOOK THAT CHANGED HIS ENGINEERING, HE NAMED ONE ABOUT A LANGUAGE ALMOST NOBODY WRITES ANYMORE. Functional Programming in Scala. The red Manning book. Scala sits at #27 on TIOBE with 0.67%. About 2.6% of developers touch it. This year’s Scala community survey drew 24% fewer responses than 2023. By every popularity metric, that’s a dead recommendation. Then he explained why he named it anyway — and it wasn’t about Scala at all: “Now I think in types.” Signature first. Implementation second. Define what goes in and what comes out, and the body becomes almost mechanical. That habit was a niche aesthetic in 2015. In 2026 it’s the highest-leverage skill in AI-assisted coding: → A 2025 study found 94% of LLM compilation errors were type-check failures. Your type system is a free reviewer for code you didn’t write → TypeScript became the most-used language on GitHub, adding over 1M contributors in 2025 to reach 2.6M. That’s not fashion, that’s teams buying a safety net → Type-constrained generation research cuts compile errors by more than half and improves repair of broken code by ~37% on average. Constraints make models better, not slower → Practical version: write the function signature before you write the prompt. Hand the agent the contract, not a paragraph of vibes → Don’t learn Scala for this. The book is a delivery vehicle for the concept. 2nd edition (2023) covers Scala 3, and the thinking ports straight to TypeScript, Rust, and Kotlin The language faded. The habit it installed became the thing that makes AI-generated code reviewable. Most people are prompting harder. The leverage is in defining the boundary before you prompt at all 👇

• 
  Asethon (@setho_asethon) reported 23 minutes ago

  @artsinni @Grxit We are not bombing countries with nuke-scale weaponry, so no. We don't. Also, you all have no problem using Windows, 30% of which is AI with large amount of gen probably getting trained on private GitHub repo's.

• 
  Bart R. McDonough (@BartMcDonough) reported 25 minutes ago

  @github I want this because the issue, session and PR are one job. Rebuilding that context across three tools is where the handoff gets slow.