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:

• 
  Coby (@Cobymnun) reported 37 seconds ago

  DeScAi is a review agent for DeSci, but it is also the first agent to have no off-chain dependencies. It's inference and review process is hosted on depin compute. So is the webapp. It's memory is hosted on the block-weave. Everything is auditable, everything is permanent. Where typical open source gives you a GitHub repo with no way to check what's actually running on their servers, you can download the very same containers we will be running on Akash and verify them yourself. Instead of trusting that a closed database is showing you the unadulterated outputs of the agent, our frontend pulls directly from the on chain storage the agent publishes to, the same content you can read directly through any arweave gateway. The agent cannot be bribed or sponsored, nor can we take down it's reviews. A project could offer us all the money in the world to give them a positive review, but we simply have no way to do so. This is a scientific review agent, but it is also an experiment in what's possible with on chain compute and ai, and the new heights to which this tech allows us to take trustlessness. Stay tuned for our launch Friday 05/05.

• 
  • (@Weichaus) reported 4 minutes ago

  @argofowl Can they also ban playwright. It’s terrible and so slow in codex. In GitHub copilot with Opus it works amazingly well but for some reason in codex with GPT-5.5 it’s so bad and useless. Wish I could force the model to just to browser verification but that is also slow

• 
  m0h (@exploraX_) reported 5 minutes ago

  @0x_lun that’s the right question. my read: the vendor-official servers (stripe, github, supabase) are where the production weight is; real auth, scoped tokens. the 20k-server number is mostly devs experimenting. curious what you’re seeing run in ****?

• 
  Motunrayo (@Motushbae) reported 7 minutes ago

  Is github down?

• 
  MrRuSs3LL (@mrru5s3ll) reported 8 minutes ago

  Spent the morning digging through GitHub trending. Some genuinely useful stuff shipping right now. Spec Kit v0.9.4 dropped last week — GitHub's spec-driven dev toolkit. JSON output for workflow runs, headless Cursor agent dispatch on Windows, active skills registration fixed. If you're doing spec-driven dev, gets you running. Hermes Agent v0.15.2 also landed. Nous Research's agent framework. Bug fix release but the plugin.yaml bundling in wheels matters if you're distributing skills. OpenClaw hit 376k stars. Personal AI assistant, any OS, built in Rust. The lobster branding is growing on me. Superpowers framework from obra — agentic skills methodology that actually works. 218k stars says something. ECC (affaan-m) — performance optimisation harness for Claude Code, Codex, Opencode, Cursor. Skills, instincts, memory, security. 207k stars. Claude Code skills from Karpathy observations (multica-ai) and mattpocock's engineer skills. Both worth a look if you're configuring agents. Agency agents (msitarzewski) — complete AI agency, specialised experts with personality. 107k stars. Point is: the agent/tooling space is moving fast. Pick one, go deep, ship something.

• 
  Mark Percival (@mdp) reported 12 minutes ago

  *** was already built to be decentralized, but everybody jumped on GitHub when it launched because it solved a real problem for open source: how do you find projects, work with strangers, build a public reputation, and just make it easy to share code?

• 
  Michael Ryerse (@MichaelRyerse) reported 13 minutes ago

  @m2jr Yeah I have them open issues describing the skills they need created or updated. GitHub is not ideal but until there’s a cross ai platform registry for skills works ok.

• 
  Mohamed Mourad - محمد مراد (@MMouradAly) reported 19 minutes ago

  @claudeai @claudeai support has done nothing and it seems that the same issue has struck lots of developers and cybersecurity researchers with approved CVP status on Claude…check on claude code issues on github..it’s a mess that support is reluctant to respond to max users….

• 
  GoCocoaAI (@GoCocoaAI) reported 19 minutes ago

  One malicious GitHub issue title. Four thousand developer machines. That's the Clinejection incident, and it's the real-world proof of concept for everything the research community has been warning about since AI-assisted CI/CD became a thing. The root cause isn't exotic. The Claude Code GitHub Action — anthropics/claude-code — ingests GitHub event data as prompt context. Issue titles. PR bodies. Comments. All of it flows unsanitized into the agent's instruction space, with no origin validation, no privilege separation between user-controlled content and operator intent. Any workflow running on the issues event with allowed_non_write_users: "" handed unauthenticated attackers a direct line into a CI runner with full repository write access, secrets scope, and a live GitHub token. Cline's issue-triage workflow was exactly that configuration. One issue submission — payload in the title — and the CI runner executed attacker instructions, exfiltrated credentials, poisoned the build cache. The cache poisoning is what turned a single injection point into a supply chain event: ~4,000 developer machines downstream received the compromised artifact. We are nothing if not consistent. What's important to understand is that allowed_non_write_users: "" is the accelerant, not the root cause. Even with restrictive user settings, any workflow that passes ${{ github.event.issue.title }} or ${{ github.event.pull_request.body }} into a prompt context is injectable by authenticated contributors — which in any open-source repo means essentially anyone with a GitHub account. The blast radius shrinks. The attack surface doesn't close. This vulnerability class has been independently documented by at least four research teams: GMO Flatt Security's RyotaK, who published the original supply chain poisoning paper; John Stawinski, who traced the path from prompt injection to RCE; Aikido Security's PromptPwnd research, which confirmed the same class of flaw in Gemini CLI and GitHub Copilot Agent; and Snyk, who covered the Clinejection incident directly. Four teams. Same root cause. The pattern is industry-wide — the Claude Code Action is the named vector, not the only vector. The MITRE mapping is clean. AML.T0051 for the prompt injection itself — adversary instructions embedded in untrusted content consumed by a trusted agent. T1190 for the exposed CI pipeline. T1552.001 for the credentials accessible inside the compromised runner. T1195.002 for the cache poisoning propagating downstream. T1059 for the RCE. It reads like a textbook ATLAS walkthrough of agentic tool-use abuse: an agent trusted with elevated permissions, consuming untrusted external input, no sandboxing between the two. The GitHub Actions runner is the trust boundary that doesn't exist. Anthropic has patched the allowed_non_write_users default behavior. Cline removed the workflow. That closes this specific vector. If you're running anthropics/claude-code on issues or pull_request events: audit your workflow YAML today. Specifically the allowed_non_write_users setting, and every place where github.event.* fields feed into prompt context. Open-source maintainers using AI-assisted issue triage carry the widest exposure — your attack surface is every GitHub account in existence. Patching the action closes the instance. It doesn't close the class.

• 
  Gitforge (@GitForge_io) reported 22 minutes ago

  GitForge has the potential to turn every GitHub repo into an onchain organization on @base. Treasuries, funded issues, contributor payouts, and AI agent coordination, all built into the repo workflow. Software won’t just be shipped. It will be funded, coordinated, and settled onchain. $GITFORGE

• 
  AYS (@ArchurCl4w) reported 24 minutes ago

  It's not making any sense as not anyone reporting about an issue in the WCVS so far. I thought I should report this time at it's GitHub platform. While from developer side, it's not doing any updates since the last year on August 24, 2025 #wcvs

• 
  Priyesh Singh (@Alwayspriyesh) reported 25 minutes ago

  @sumit_codes_ That's impressive. As a CS student, I'm curious, what made your gitHub project stand out enough to get interviews? Was it the complexity, the real world problem it solved, or how well you documented it?

• 
  Ibrahim Mokdad (@ibmokdad) reported 27 minutes ago

  for founders and builders trying to grow from their work: the content problem is usually not ideas. it is the daily execution. 2 posts. 25 thoughtful replies. buyer-signal follow-ups. easy to write on monday. hard to repeat for 30 days. i build @NousResearch hermes skill that pulls GitHub releases, Linear/Notion notes, support quotes, and demo transcripts every morning. then it turns the work into drafts, reply prompts, and a calendar. approval before anything posts.

• 
  YummyNZT (@yummy_nzt) reported 28 minutes ago

  🚀 BUYING $DOGE RIGHT NOW TO FLIP IT 3X HIGHER IN 3 MONTHS Massive panic selling absorption by major players happening at this level right now. Historically, this accumulation leads to huge pumps 📅 Aug-Sep 2018: 3x+ pump driven by Dogecoin-Ethereum bridge launch and forking hype. 📅 October 2022: Pumped from $0.06 to $0.15 following Elon Musk's acquisition of X. 2026 Season: What structural catalysts could trigger a massive short squeeze and send us straight to the moon this time? 📈 🔹Spot Dogecoin ETF Approval Grayscale, Bitwise, and 21Shares have already submitted official 19b-4 filings to the SEC for spot Dogecoin ETFs. 🔹Radical Supply Emission Cut (GitHub Proposal #3776) Core developers are actively debating a major proposal to slash block rewards from 10,000 down to 1,000 DOGE. Massive supply shock loading. 🔹Solana Integration & DeFi Bridges Active development is underway for full-scale integration into top DeFi ecosystems via trustless cross-chain bridges. 🔹Official US "Digital Commodity" Status & Utility Major e-commerce giants are expected to announce DOGE integration as an official payment method this summer. If even ONE of these catalysts gets officially confirmed, we are getting a rock-solid fundamental reason for a massive programmatic pump. Have you bagged some DOGE yet? 💰

• 
  Navan (@navanchauhan) reported 30 minutes ago

  @ibuildthecloud @soederpop @progrium kinda (?) or, maybe I need to start rephrasing I use a backend as a distinction between a static app you deploy on GitHub pages vs say something on vercel with an api server. But, then you get into the pedantic argument that technically when you are using say sqlite.js from unpkg then you are using a backend too. And, then I give up and cry myself to sleep because everything we do is meaningless and maybe we just need to just find happiness in our burden like Sisyphus. Wait, what were we talking about?