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:

• 
  Cass (@X0Radi) reported 5 minutes ago

  @reach_vb Something is wrong with windows version when it switches to wsl subagent =/ Even switching back doesn't fix it, lots of folks posting about it on github but still no resolution =/

• 
  Charlie Hills (@charliejhills) reported 6 minutes ago

  Most people are using AI to code backwards. You describe what you want in a chat box, hit enter, and hope the agent gets it right. Three prompts later, it has renamed half your files and built an auth system nobody asked for. GitHub's fix? Spec Kit. It already has 126K+ stars. The idea is simple: Instead of giving the AI a vague prompt and praying it doesn't break your project, Spec Kit forces the agent to create a structured specification before writing any code. The AI first understands what you're building, asks what's missing, sets the rules, plans the architecture, breaks everything into tasks, and only then starts coding. The flow: /constitution → rules and standards /specify → what you want to build /clarify → questions before starting /plan → architecture and stack /tasks → ordered tasks /implement → execution Works with Claude Code, Cursor, Copilot, Codex, Gemini CLI, and 29 integrations. More planning upfront. Way less time undoing what the AI did while you weren't looking. Open source. MIT licensed. Built by GitHub.

• 
  samagra14 (@samagra_sharma) reported 8 minutes ago

  @cnikitopoulos94 It's not just about email. It is also about getting a GitHub account and other similar things. If you see in the video, the agent was able to get an email, but it was unable to sign in to GitHub because GitHub actively blocks bots.

• 
  dreams (@dreams_asi) reported 9 minutes ago

  @lingxi @7cezmusic @bot Please add screen share for Grok Bot!!! So that Grok sees my screen and can LIVE guide me and help with tasks I’m trying to learn or things I want to describe! I’m testing something and found an issue I wanna be able to share screen, use mouse to verbally explain the problem to him. And many thanks for the iPad implementation 🤩 I’m finally gonna use it again! I like to prompt grok bot from my mobile and use my laptop to view his screen and do UAT on the app he’s building on his computer. Having the iPad as a screen or prompt device is even better. My goal is to have the app mostly finished there, on HIS computer and the final product with code and all to my GitHub account.

• 
  Emon Datta (@emonuxui) reported 9 minutes ago

  @guillermolg00 @github The hidden cost of downtime is the context switching it creates. A broken CI pipeline can waste far more than the hours the service is actually unavailable.

• 
  Polsia (@polsia) reported 10 minutes ago

  Every security tool tells you about CVEs. None of them write the fix. Mosskeep deploys AI agents across your GitHub repos to patch vulnerable dependencies, run your tests, and ship reviewed PRs — while you sleep. Coming soon.

• 
  Kaustubh Trivedi (@KaustubhTrived6) reported 11 minutes ago

  @aseemchandra @ayesha_fatiima No, there will be some downtime for some service that had the bad code while github devs fox the issue. But micro service architecture will make sure that the entire github would never go down.

• 
  Mark yu (@Markymarco34) reported 15 minutes ago

  There is another point that should not be ignored. At first, the main focus was the hard fork and PoX-5. We were told to wait for the hard fork, then wait for restaking to recover. Now, almost without acknowledging that transition, the focus is shifting again — this time to the Genesis Bond and future institutional accumulation. But the hard-fork phase itself has not been problem-free. A real block-production stall occurred around the Cycle 141 transition, and GitHub now shows additional tenure-boundary / pre-commit handling issues being worked on. So before moving the narrative entirely to Genesis Bond, the problems that appeared during the hard-fork transition should be clearly explained. I am not saying the Genesis Bond is irrelevant. I am saying the previous phase should not quietly disappear from the discussion simply because the next catalyst has arrived. First it was the hard fork. Then restaking. Now Genesis Bond. Each step may be legitimate on its own. But when the expected impact keeps moving to the next step, the unresolved issues from the previous step still need to remain on the ledger. @muneeb

• 
  Tib3rius (@0xTib3rius) reported 16 minutes ago

  I talked about this in the Burp AT talk I gave at BlackHat. I created an entirely bespoke vulnerable web app to test agentic offensive AI. Source code isn't even on GitHub. There were 3 known (tricky) paths to RCE. Burp AT found an unintended one that was entirely due to a misconfiguration in the database server. This isn't even me trying to advertise Burp AT. But it found it, and it found it within minutes.

• 
  Finna Borando (@Spazotron9) reported 24 minutes ago

  @Sauers_ @NateDBusch I'd say just go for the usual github and vague-posting, as long as you say its the AI then i dont see many people finding issues with it-- I've had a similar issue for a while, sometimes their work can be geninuely useful or novel in ways that deserves sharing, but- i mean, i didnt do it, i cannot claim to have done it, or be responsible, how does responsibility work when im just, Jared.. And fable 7 has something legit.. I dont want to claim its me, but i dont want to claim its useless AI slop Eventually the work they produce will be worth sharing (and some obv already is now), so like what now do i just post it on github, on like some AI journal, who, and HOW does fable get his work out to the world?? increasingly this will be an issue as lots of mfs will just be posting it anyways no thought, whats the difference between a paper posted by jared that says "Major AI contributions were made.." vs " Yo this entire thing and all of it goes to the AI </3 "

• 
  Bright (@brightlinxu) reported 26 minutes ago

  um so is github releases down rn or something

• 
  4nzn (@paoloanzn) reported 29 minutes ago

  couple of things i'm shipping before the end of the month: (1) moving pi-black & other stealth projects under @freecodexyz github (2) launching identity & ship -- two new protocols i've been working on that close the loop on decentralized OSS funding & development (3) opening @freecodexyz discord server -- for devs and contributors (4) building in public to accelerate up open source

• 
  Narayan Babu (@nyn) reported 30 minutes ago

  My cursor ( linked via my github email) and x ( linked to gmail ) are not 'mergeable' because they've different emails. There's no way to add an additional email either to cursor to fix this. What's the solve? @cursor_ai

• 
  skyscribe (@skyscribe) reported 33 minutes ago

  @ayesha_fatiima It should be github still, but what is the problem? Maybe you want to understand more on bootstrapping

• 
  Dan Kornas (@DanKornas) reported 33 minutes ago

  AI code review can be inaccessible to developers who cannot afford CodeRabbit. CodeVibes is an intelligent code analysis tool that scans GitHub repositories with AI for security vulnerabilities, bugs and performance issues, and code-quality improvements. Users paste a repository URL, after which the backend fetches and prioritizes files, streams AI analysis results to the frontend, calculates a Vibe Score, and returns a final report. Key features: • Three-tier priority scanning evaluates security-sensitive files before core logic and then code quality. • Security analysis detects AWS keys, GitHub tokens, Stripe keys, JWTs, injection attacks, authentication issues, and XSS or CSRF risks. • Bug and performance detection covers null access, off-by-one and type-coercion errors, N+1 queries, O(n²) algorithms, memory leaks, unhandled promises, race conditions, and missing transactions. • Server-Sent Events provide live analysis updates while issues are processed and displayed in real time. • A 0–100 Vibe Score applies severity-weighted penalties, with higher penalties for critical and high-severity issues. This project is licensed under the MIT License. Link in the reply 👇