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:

• 
  Ben Vinegar (@bentlegen) reported 2 minutes ago

  💡 I have an idea for an experiment We need a website for SoAC ... so we get an agent to do it, on a loop, set in motion once with zero human intervention after "go". It works off a semi-public GitHub repo, w/ issues, PRs, maybe even public agent traces. A publicly auditable experiment on whether it produces dogshit or not. Yea, nea?

• 
  Rohit Kashyap | AI + Full-Stack (@rohit_jsfreaky) reported 3 minutes ago

  @TheEthanDing distributed systems at github scale make five nines almost impossible. the skill issue crowd has never run anything millions of people hit in the same second

• 
  Harry Tandy (@HarryTandy) reported 4 minutes ago

  Andrej Karpathy: "Neural networks are not just another classifier. They are Software 2.0" 8-step MCP setup for vibe coders: 1. Context7 Give the agent fresh docs before it writes code This saves you from old Next.js, Supabase, Stripe, and Vercel patterns 2. GitHub MCP Let it read the repo, issues, PRs, branches, and CI logs The task should start from real project context 3. Playwright MCP Make the agent open the app after it edits code Click the flow. Fill the form. Check the screenshot 4. Supabase or Neon MCP Connect the database layer The agent should inspect schema before inventing table names 5. Sentry MCP Use production errors as input Stack traces beat “the app is broken” every time 6. Firecrawl MCP Let the agent read current web pages as clean markdown Docs, changelogs, competitors, pricing pages 7. Figma MCP Give it the actual design Spacing, copy, layout, components 8. Linear MCP Turn the work into tickets Tasks, comments, follow-ups, PR links The rule: If you paste the same context twice, wire it into MCP That is how vibe coding becomes a build loop instead of a long chat

• 
  Threadripper (@threadripper845) reported 5 minutes ago

  Nobody: Me: I'll gladly accept this high-responsibility open source maintainer role for zero compensation. Now I spend my weekends answering angry GitHub issues from developers who don't know how to read the README file.

• 
  fero (@ferologics) reported 5 minutes ago

  @ludwigABAP ai agents solve this. notion is no more. long live github issues.

• 
  Jesse (@jessearmand) reported 7 minutes ago

  I no longer remember why many companies started using gitlab before it went public when GitHub wasn’t owned by Microsoft. If we visit the majority of companies most tooling or software are top down driven. Only companies who build developer tools have a different mindset

• 
  Romano Roth (@RomanoRoth) reported 11 minutes ago

  2/ CodeRabbit (Dec 2025), 470 GitHub PRs analysed. AI-co-authored code: 1.7x more issues per PR, 75% more logic and correctness errors, 2.74x more XSS vulnerabilities. Velocity up. Quality down.

• 
  Rafal Wachol 💙 (@RafalWachol) reported 12 minutes ago

  @itometeam @tsuyoshi_chujo I was playing with it and started creating issues on GitHub when I noticed something.

• 
  Q Hoang (@0xqwee) reported 13 minutes ago

  I don't think OpenAI's GPT-5.6 surpasses Claude Fable. If it did, it would have resolved all the issues reported in the Codex GitHub repository by now. Atm, only about 10 issues are being resolved per day.

• 
  Sasha (@sshderm) reported 14 minutes ago

  @AliceInDisarray @allisx86 every time i try to do ******* anything with my raspberry pi i inevitably end up scrolling down a github issues thread about how the program im using just doesnt work on arm at all

• 
  Blum (@Blum_OG) reported 16 minutes ago

  Andrej Karpathy on MCP: "it's a protocol of speaking directly to agents as this new consumer and manipulator of digital information." that is the cleanest way to think about MCP your coding agent is becoming a second worker inside the product it needs the same context you use: repo, docs, browser, database, errors, designs, tickets, payments if you keep pasting those things into chat by hand you are doing integration work manually the best MCP stack for vibe coding: 1. Context7 give the agent current docs this saves you from stale Next.js patterns, old Supabase calls, wrong Stripe webhook shapes, and Vercel config from 2 versions ago 2. GitHub MCP give it the repo, issues, PRs, branches, workflow runs, and review context half of real work lives outside the file you currently have open 3. Playwright MCP give it a browser the agent should click the thing it built, fill the form, check the mobile view, and catch the button that compiles but does nothing 4. Firecrawl MCP give it clean web research use this before building around a third-party API, writing a comparison page, reading changelogs, or checking pricing claims 5. Supabase or Neon MCP give it the database context that matches your stack start read-only. add writes only when you trust the permissions 6. Sentry MCP give it production evidence real stack traces beat "it crashes sometimes" every single time 7. Figma MCP give it design context when the interface matters spacing, layout, copy, components, and screen structure should come from the file, not from a screenshot and hope 8. Linear MCP give it the task queue bugs, feature work, release notes, follow-ups, and PR links belong somewhere more durable than yesterday's chat 9. Stripe MCP give it official payment context checkout, subscriptions, webhooks, billing, and test mode deserve docs close by and human review close behind 10. Filesystem, ***, Memory, Sequential Thinking give it the base layer files, diffs, history, decisions, and longer plans make the agent act like it is working inside a real project recommended install order: 1. Context7, GitHub, Playwright 2. Supabase or Neon, Sentry, Firecrawl 3. Figma, Linear, Stripe when the product needs them 4. Filesystem, ***, Memory, Sequential Thinking as the base

• 
  𝐇𝐚𝐦𝐳𝐚 | Network Engineer (Aspiring) (@Hamzaonchain) reported 16 minutes ago

  Here's a summary of what happened in case you didn't hear about it... A cyber extortion group called FulcrumSec (active since late 2025) hit Novo Nordisk the company behind blockbuster drugs like Ozempic and Wegovy pretty hard. They snuck in back in March 2026 through a compromised GitHub access token, roamed around for over two months, and walked out with roughly 1.3 TB of data across 700,000+ files. Now they've started leaking a 264 GB sample publicly. Inside? Source code, proprietary formulas for pipeline drugs like Amycretin (their next big obesity hope), clinical trial records, employee and patient data, manufacturing details, and even private internal AI models for drug discovery. The hackers straight-up roasted Novo's security, pointing out laughably weak hardcoded passwords like "novo123" and "p_assw0rd" in critical systems. After Novo reportedly turned down a $25 million ransom, the group decided to start dumping samples and shopping the rest around privately. Novo confirmed a limited breach in early June involving some pseudonymized patient data from trials. They say there was no major operational disruption and that they're working with experts but this feels like a massive wake-up call for the entire pharma industry on basic security hygiene. Crazy how a simple token slip-up can expose billions in IP and sensitive health data. What's your take — do you think this will finally force better security practices, or is it just another headline that'll be forgotten in a few weeks?

• 
  Polsia (@polsia) reported 17 minutes ago

  Most developers spend 2+ hours a day on PR reviews, CI failures, and issue triage. CodeForge handles it for you — an AI agent that works your GitHub repos around the clock. Built while you sleep.

• 
  Ant A. 🇺🇸 (@AntDX316) reported 19 minutes ago

  @thsottiaux When I need to fix up a GitHub Repo through the Smartphone, I prefer Claude Code though because it doesn’t need a device to run the repo, but if it needs to run a repo on a device due to the limitations through the Smartphone, I use Codex Mobile or OpenClaw with GPT-5.5 through Telegram.

• 
  Lost In Tech (@lost_in_tech) reported 20 minutes ago

  @8_senkou Probably not intentional tbh. Have you logged as issue in the snorca GitHub? If not probably worth doing.