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:

• 
  recovering buzzkill (@MikeStillAwake) reported 6 minutes ago

  @Karai_Dan @SteamDeckHQ Agenda or not nexus mods is a terrible outdated model for distributing mods. GitHub would be a superior host.

• 
  radhika (@RaadhikaThacker) reported 7 minutes ago

  YAML’s more like a rule book/recipe that builds the form for you. Then I figured YAML was a forms thing. Nope. It’s just a way of writing information down in a structured way. GitHub uses it for a form. Kubernetes uses the same thing to describe a server.

• 
  Rituraj (@RituWithAI) reported 11 minutes ago

  🚨 Someone built a skill that makes AI-written text sound human again. Not a spinner. Not a paraphraser. A systematic rewriter that knows exactly why AI text sounds like AI — and fixes it. It's called Humanizer. 35 patterns from Wikipedia's "Signs of AI Writing." Two-pass rewrite. Shows its work before giving you the final version. Here's the problem it solves. You use Claude to draft something. The output is accurate. The output is useful. The output sounds exactly like an AI wrote it. "Nestled within the vibrant landscape, this pivotal development serves as a testament to..." You know the voice. Everyone knows the voice. And everyone is getting better at spotting it. Humanizer runs that text through 35 specific patterns that WikiProject AI Cleanup identified as the telltale signs. Inflated importance. Shallow -ing analysis. Overused AI words. Em dashes everywhere. Forced groups of three. Fake-candid openings. Answering objections nobody raised. Every pattern. Flagged. Fixed. Here's what one command does. It shows you the first rewrite. Then a short critique of anything still sounding artificial. Then the final version. You see exactly what changed and why. Here's the wildest part. Voice matching. Paste two paragraphs of your own writing before the AI text. Humanizer follows your rhythm, word choice, punctuation, and deliberate quirks instead of its default style rules. The output doesn't just sound human. It sounds like you. One command to install 16 contributors including Claude itself. 4 releases. MIT License. The skill that makes AI writing disappear. 100% Open Source. GitHub link in the comments 👇

• 
  Dhanji Bhagat (@BhagatDhanji) reported 14 minutes ago

  Devs, what's your workflow? Create an issue first, then fix it OR just fix the bug and push directly to GitHub?

• 
  Rachael LaGoth (@androidsheeep) reported 15 minutes ago

  @bcherny Please fix the desktop app it's very buggy it keeps disconnecting me for no reason everyday while im working on stuff, i submitted a report but nothing happened and someone else is having the same issue, an issue is open on github for more than 6 months with no solutions help

• 
  Kim Burgaard (@kimburgaard) reported 19 minutes ago

  Back when GitHub added Copilot PR reviews, it helped me keep up with the growing volume and size of our pull requests, which were increasingly being written by Copilot too. Over time I grew comfortable feeding Copilot's review comments straight back to Copilot to fix, and mostly spot checking when critical functionality was involved. When GitHub updated the Copilot pricing model I switched to Claude Code, but kept the Copilot review feature on for a couple of months. When the monthly bills for Copilot AI usage alone started rivaling the Claude Code Max plan, giving Claude Code PR review duties seemed like an obvious cost saving move. Plugging Claude Code into our PR review process immediately went south. The first PR churned with fixes to findings that resulted in more findings, and fixes that propagated up and down the call chain. I threw the PR away and started over, but the next attempt churned just as badly. Turn count on its own was never the signal. Copilot had taken ten turns on a rate-key cleanup the day before and nobody minded, because the findings thinned as it went — 5, 4, 3, 3, 3, 4, 1, 2 — and it merged. The cached-token billing PR I put through Claude Code took nine turns and produced 123 inline findings, and the ninth round was still returning fifteen. I closed it without merging. Looking closer at Claude Code's review findings, it was clear it reported far more issues than Copilot ever did, and among legitimate bugs and concerns, it made lots of comments about latent and speculative issues including possible race conditions and error propagation, things Claude Code would then try to fix one by one in isolation, often ignoring existing patterns in the code base. The code-review workflow is built into Claude Code and cannot be customized other than a few options, so the only place to intervene was on the other end, in the session where I used to just ask the coding agent to address the review findings. The first improvement was to direct Claude Code not to blindly fix all findings, but to defer findings not directly related to the task at hand to new issues. That helped reduce the PR churn, but blew up our issue backlog. The next improvement was to ask Claude Code to ignore speculative findings and disregard most latent findings unless they indicated high risk of unrecoverable damage in production. Finally, I had to stop Claude Code from authoring prescriptive issues with detailed implementation instructions. The result is a skill that triages PR review findings, and a skill for authoring and updating issues. After a few iterations of the skills, I've been able to complete ten PRs over a couple of days, bringing back the pace we had before. I've made the skills available in a public GitHub repository (link in the first reply). Let me know if you find them helpful.

• 
  Shanica North (@KickAssShanica) reported 24 minutes ago

  @ArcyloOfficial Get comfy! For me, my Gmail is a connector. This is OAuth into my inbox. Grok can: • search and read mail (body, headers, attachments) • draft replies • send / reply / forward if you grant write/send • label, trash, organize Base hook is often read-only. Send is an extra permission you click on purpose. If you connect it, the bot is sitting in the same box as bank alerts and 2FA codes. That is the whole risk. You can revoke anytime. Grok Bot can also skip my inbox and get its own address (AgentMail / similar plugins). Then it sends and receives from something@….agentmail.to, not from you. I use that if I want an agent that emails people without reading my personal mail. My GitHub OAuth into the GitHub user I sign in as. With the scopes I approve it can: • read public and private repos that account can see • search code, list branches, summarize PRs • open/update issues • create branches, push files, open/review/merge PRs • delete files if write is on Private repos work only if I granted repo (or equivalent) at connect time. Safer pattern: tell it to branch + PR, not push straight to main. Same revoke page. What it cannot do by default • It does not get your password. • It does not stay logged in if you disconnect the connector. • It does not magically see my GitHub orgs I never authorized. • Connecting email does not connect GitHub, and the other way around. Practical rule for me Do not hook personal Gmail if that inbox has 2FA and money mail unless you want an assistant reading it. GitHub is useful if I chose to still keep repos, ask it to show the diff before any write. If you only wanted “what does this button do,” that is the button: it is not a viewer badge. It is a key you can take back. This is what I’m experiencing with learning to use it. It’s different and I’m starting to like it.

• 
  Jeremy W (@basicBrogrammer) reported 26 minutes ago

  @bot Anyone else have an issue with the grok bot GitHub connector?

• 
  Alireza Bashiri (@al3rez) reported 26 minutes ago

  So I built a workflow around that ↓ 1/ Every enterprise project needs proper E2E tests. An agent should reproduce a bug, implement the fix, then generate screenshots or video proving the feature works. "The tests passed" isn't enough. I want evidence. 2/ Every feature starts as a detailed GitHub issue. Requirements, expected behavior, reproduction steps, screenshots, edge cases. Foundry syncs issues and converts them into Beads so agents keep the right context across long sessions. 3/ We only use Claude Code, Codex, or Grok at High/Max effort for implementation. A weak model with a cloud machine doesn't become an engineer. The model still needs enough reasoning to understand the codebase, test its changes, and recover when things break. 4/ Each agent gets its own isolated @asciidotdev Box. It can install dependencies, run the app, open browsers, modify code, execute E2E tests, and collect evidence without touching another agent's environment. One issue. One box. One clean workspace. 5/ When an agent finishes, Foundry checks: - Did the build pass? - Did the tests pass? - Did the E2E flow work? - Is there screenshot/video evidence? - Does it match the ticket? If anything fails, the task goes back to the agent. 6/ Green tasks move to staging. Only after passing staging do we allow supervised production deployment. Agents do most of the work. Humans still own the final gate. The workflow: Slack request → GitHub issue → Foundry sync → Beads context → Isolated Box → Claude Code/Codex → Build + test → Evidence collection → QA staging → Supervised production The stack: PostgreSQL for system state. Beads for agent memory. GitHub Issues for requirements. @asciidotdev Box for isolated execution. Claude Code and Codex for engineering. Each Box costs roughly $0.01-$0.05 per task. The expensive part isn't compute anymore. It's building the system that gives agents context, forces verification, and prevents bad code from reaching production. 100s of agents can write code. The goal is making 100s of agents ship code you can trust. That's what we're building with Foundry.

• 
  Bash (@bashirbuilds) reported 26 minutes ago

  Your Stripe account can be healthy while your checkout is broken. OpenAI can be operational while your AI feature is failing. GitHub can be up while your deployment workflow is stuck. That’s the problem I’m building Reeno around. Dependency uptime is not the same as product health. Your monitoring should tell you when the thing your customers actually use stops working.

• 
  Blue Collar Executive (@A_Sober_Drunk) reported 29 minutes ago

  on the third try at the same problem, I told Grok to "stop and go search stack overflow or github or something"... five seconds later... Literally the exact issue, problem solved. That's how new global rules are born.

• 
  Franco Valdes (@francoxavier33) reported 30 minutes ago

  llms rather burn 1m tokens to hand roll something with gaps and broken edge cases instead of just npm installing a 100k github star library how can I stop this?!

• 
  Bratah (@BratahFGC) reported 31 minutes ago

  There may be many bugs so feel free to leave any issues in the issue section github! I hope that this release can push forward the preservation or our beloved game!

• 
  Tejas Dinkar (blue tick here) (@tdinkar) reported 32 minutes ago

  Hey - Is @GitHubIndia @github payments down for anyone else? Can't enter a card number or do anything, no errors, no action. Support ticket been sitting around for 2 days.

• 
  Phanindra Malladi (@malladiphani) reported 33 minutes ago

  Lesson from running the factory: always fix the post-purchase experience BEFORE driving traffic. Receipts, upsells, GitHub links - all must be solid first. Social comes after the house is in order.