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:

• 
  Alex Metelli (@ametel01) reported 9 minutes ago

  @GergelyOrosz Yes this is becoming a problem but I found there is kind of a smell to this kind of offers - they always do the first reach out - they mention some of your GitHub repo in a shallow way and tell you are a great fit for xyz - they start by asking you AI generated technical questions in chat - and of course finally they invite you to malware ridden repos they want you to run I’ve asked LLM to analyse one of this repo once and it found an hidden python script in one of the deps that would take over my webcam. Definitely BE PARANOID

• 
  Miro Jomaa (@MiroJomaa) reported 10 minutes ago

  @vboykis Problem is still, and you saw that with the sandbox espace at openai, that if one agent figured a "hack", that all the other agents see it as an internal message and follow it. Thats a big alignment problem. 1 agent does ****, all follow. Like abusing json, github actions, dto.

• 
  Aditya Roushan (@Adityarou6688) reported 11 minutes ago

  Most AI models are completely blind to the live web. If you ask your assistant to read a Reddit thread, check a trending discussion on X, or pull comments from a YouTube video, it usually hits a wall. You get blocked by bot detectors, walled gardens, or expensive paywalls for every single platform. To fix this, you’d typically need to buy separate API access for a dozen different services and wire them up manually. That just changed. An open-source project called AgentReach just hit 38,000 GitHub stars by solving this exact headache. It gives your AI local read and search access across 13 major platforms (including Reddit, X, YouTube, GitHub, and Instagram) through one tool, with zero API fees. Here is why it works: Built-in proxy support to get around hard IP blocks Scrapes bot-resistant and unauthenticated pages cleanly Pulls live data and transcripts straight into your AI coding setup You run a single command in your terminal, and your AI can suddenly search real-time public conversations across the web.

• 
  DarkestToaster (@DarkestToaster) reported 14 minutes ago

  Your downloads don't work proper cause IA is not a reliable website and it crashes a lot, also your Github torrent or file download is broken it goes to a broken link, upload this somewhere else and just provide all the files in the .rar / .zip next time cause this sucked to see simply because I didn't get pushed this post on Twitter 2 weeks ago now I'm **** out of luck to play it, as gonna stream it too but.. Oh well

• 
  Pascal Bornet (@pascal_bornet) reported 19 minutes ago

  Even fast-food chatbots are preparing for technical interviews. You open a support chat because you want lunch, but somehow the conversation ends with the chatbot solving a Python problem that has nothing to do with burritos. I have to respect the ambition. Pepper is supposed to answer questions about orders, but it is clearly using the job to build experience before applying to a software company. At this rate, the next time your order is late, please be patient because the chatbot may be updating its GitHub portfolio or completing the final round of a coding interview. Which customer-service chatbot secretly has the most promising tech career? #AIHumor #WorkplaceHumor #AgenticAI

• 
  Chris W (@Chris_Wozniczek) reported 22 minutes ago

  Me in the past: > open plan mode and send a detailed planning prompt > approve plan and start implementation > get result - check - send prompt to fix bugs - check > send a new prompt to double check all the implementation plan is done > find tasks not finalized ask to finalize > check and test if that works > finds ui and ux issues - ask to fix - review fix > push to github > restart with new feature Me now: > send an idea what i want to build > review the plan > close laptop [groceries, kids, hanging out, consulting, kids, doing life in the meantime] - come back and to review ready PRs and watch testing video from ui and functionalities testing - click merge done

• 
  Valentyn Kit 🦀 | Rust · Solana (@valentynkit) reported 23 minutes ago

  @tsoding the reliable part of github is that it goes down predictably enough to farm content off it

• 
  Frank Hale (@h7software) reported 24 minutes ago

  @edandersen Why didn't this guy have a backup of his work? I would have started there before crashing out on Github Issues.

• 
  Tolani (@Tola_niii) reported 30 minutes ago

  @n_ev_er_mi_nd @commonsmade scroll down X is below github The reason why I received zero point is because you are yet to link them Sadly the people you vouch for earlier before connecting socials will remain zero 😭

• 
  AGTP (@AGTPinsights) reported 31 minutes ago

  Anthropic just rolled out Claude Mythos 5 for enterprise code security scans today. Here's what you need to know. Claude Security scans now run on Mythos 5, Anthropic's most capable cybersecurity model, in public beta for all Claude Enterprise customers starting August 21, 2026. Enterprises can use it without needing separate access to Mythos 5 itself. Point Claude Security at a GitHub repo and Mythos 5 traces data across files, reasons about how components interact, and flags vulnerabilities. Each finding includes a CWE category, confidence rating, severity rating, and a suggested fix, and patches open in Claude Code on the web. Mythos 5 itself stays access-restricted, running only behind the scan, so it returns findings without exposing the underlying model directly. This is the first time Mythos 5 capability has reached anyone outside the original Project Glasswing partner group. Alongside this, Anthropic is integrating Mythos 5 into partner security products and launching a Defender Advantage Fund with $35 million in Claude credits for open-source security work. Key numbers: - $35 million in credits for the Defender Advantage Fund - Mythos 5 pricing: $10 per million input tokens, $50 per million output tokens - Over 10,000 high- or critical-severity vulnerabilities found via Project Glasswing as of May 2026 - ~50 initial Glasswing partners, expanded to ~150 organizations in 15+ countries by June 2026 Claude Mythos 5 launched June 9, 2026 and remains limited to vetted partners for direct dual-use access.

• 
  divyansh tiwari (@DivyanshT91162) reported 34 minutes ago

  THIS CHANGES HOW WE THINK ABOUT AI TEAMS. Jack Dorsey, Twitter’s co-founder and former CEO, just released an open-source framework designed to let AI agents operate like actual members of a company. Not just a chatbot. AI agents can be added to channels, given permissions, connected to tools, and made to collaborate with humans and other agents in real time. And the wild part? The entire thing is FREE and already has 29,000+ GitHub stars. Setup takes roughly 5 minutes: 1. Clone the repository. 2. Deploy your own server. 3. Connect channels, search, ***, and automations. 4. Add your AI agent like a new teammate. 5. Give it the right permissions and let it work. This is getting dangerously close to the idea of an AI-native company where software agents handle entire workflows instead of simply assisting humans. This isn’t just another AI tool. It’s infrastructure for building AI-powered teams. Repo👇

• 
  at drath vader zero zero zero (@drathvader000) reported 35 minutes ago

  @vaxryy Atp get a broken laptop and self-host your own *** server, you'll have far better uptime than github

• 
  max guy 😐 (@GolerGkA) reported 38 minutes ago

  @BenjDicken @davidfowl The main thing a GitHub agent-first competitor should do is just to cut down on features. Really hard. Everything to get performance and reliability.

• 
  Colin O'Brien (@InsanityBit) reported 38 minutes ago

  Things that won't change supply chain security meaningfully: 1. "Don't use package repositories" Attackers will just distribute via github or whatever your defacto platform becomes. 2. "Don't support build scripts" Attackers will use `./configure` scripts, makefiles, etc. 3. "Have a bigger stdlib" Literally will do nothing, python's stdlib is famously large and attacks have happened. Worst suggestion ever btw. 4. "Don't install new dependencies, use cooldowns" Attackers will yank versions or issue critical CVEs. This is the SMS 2FA of supply chain mitigations. 5. "Audit your dependencies" Developers are not in a position to do this and tooling does not make it easy. AI helps though so this is by far the best suggestion, but we need better tooling support.

• 
  Phillip Shoemaker (@pbsIdentity) reported 41 minutes ago

  India just ordered hundreds of Google Firebase accounts shut down after authorities found scammers using the platform to impersonate major banks. At least 57 Firebase-hosted websites and databases were targeted for takedown this month alone. Some mimicked banks. Others distributed malicious Android apps. Some were designed to steal financial information from phones. Here's what I find interesting. Firebase isn't some shady hosting company operating out of a basement. It's Google infrastructure. That's exactly why criminals want it. We've spent years teaching people to look for obvious signs of scams. Weird domain. Broken English. Sketchy hosting. Browser warning. No HTTPS. But increasingly the attacker doesn't need to build suspicious-looking infrastructure. They borrow legitimate infrastructure. Google. Microsoft. Cloudflare. GitHub. Dropbox. Whatever gives the attack credibility and reliability. Now imagine the average person inspecting the link. They recognize Google. The connection is encrypted. The page loads perfectly. The certificate is valid. Everything their brain has been trained to interpret as: SAFE may technically be true. Except the person controlling the page is a criminal. That's an important distinction. HTTPS proves your connection to the website is encrypted. It does not prove the person operating the website is honest. A Google URL proves Google is providing infrastructure. It doesn't necessarily prove Google created the content you're looking at. The little padlock was never a morality detector. We just accidentally trained an entire generation to treat it like one. India says scammers have increasingly shifted toward Firebase because its legitimate development tools and database functionality make it useful infrastructure for fraudulent sites and apps.