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:

• 
  Azael (@theazaelov) reported 1 minute ago

  This guy makes $20,000-25,000 a month from a dashboard where AI agents push code run social media and close deals for him. No in-house developer. No social media manager. No project manager. At the core is a multi-agent dashboard. It is not one chat with a model. It is a distributed command center where every agent has its own role. Jarvis is the squad lead. Forge writes code. Ghost handles content and SEO. Hype runs social media. Scout digs through research. Closer drives outreach and sales. The whole stack costs $200-800 a month. His command center separates strategy from execution. The owner keeps the strategy. The agents take over the execution entirely. What he does: → The owner sets a goal. Jarvis breaks it into tasks and spreads the cards across the Kanban board between the agents. → Forge creates branches in GitHub opens a PR and ships a site or an MVP. → Ghost writes SEO articles and a changelog. Hype prepares posts and a cadence for the X accounts and fixes broken-image posts. → Scout digs through competitors and new opportunities. Closer collects leads and sends out applications through the outreach pipeline. → Every status flows back into Mission Control. You see what is done what is in review what is stuck and which agent is responsible. → The agents are tied to real assets: Shoa Dev, Moltza, AI Tools Directory, Vydra, ClawHub. Code terminal browser and pricing page are open right next to them. I broke the economics of this command center down into three scenarios. Results: Replacing the team. A manual lineup of a developer an SEO copywriter an SMM a researcher an outreach specialist and a project manager costs $12,000-25,000 a month. The whole AI stack is $200-800. Even at 50-60% of a live team's quality that is $10,000-20,000 saved a month. The setup product. The Shoaf Systems pricing page has three tiers: $199, $499 and $1,499. 20 quick setups 10 business setups and 5 full systems a month bring around $16,465 in revenue and $10,000-13,000 net. The retainer. Support runs $499-1,500 a month. 15 clients at $799 give $11,985 MRR. Together with setup that is $20,000-25,000 a month or $650-830 a day for one operator. If you have ever run several projects at once you know this friction. Either you hire a team for every function and pay from $12,000 a month. Or you tear yourself apart between code content and sales. Or you take on contractors and drown in approvals statuses and blown deadlines. Now he deploys one Mission Control. It unfolds into a matrix of 6 agent roles 7 types of business and 5 ready departments on the marketplace. That is more than 200 configurations of AI teams for any niche without a single new hire. For everyone building an agency a SaaS or a service company one thing matters. A business does not need a smart answer from a chat. It needs constant forward motion of tasks. One chat closes a question. A command center of agents closes a process. At the same time the real power is not in the agents and not in a pretty dashboard. It is in the operational memory that stores decisions texts leads and mistakes. A live team starts from a blank page every time. The system gets sharper with every cycle. In the end he turns an entire operations department into a set of roles and prompts. A new business function is no longer a reason to hire a person. It is just one more card on the board.

• 
  Vatsal | Wishlist SHOCKFACE 🔌 (@VatsalAmbastha) reported 5 minutes ago

  @MjTheHunter Some ffmpeg based targa/png to jpg converters without opening PS. I just type a command and it batch converts them with configurable quality levels We rely on 3rd party assets a lot and they often have 4k/8k tex, but I'm ok with jpg quality usually. I didn't want tar files to bloat up my repo and end up paying GitHub for additional storage every month. Also, I've run into issues many times because a tar file was 100mb+ and GitHub won't take it at all! Then I have to setup *** LFS and I hate that. </rant>

• 
  LetsGo (@askperp) reported 5 minutes ago

  What happened to Heavy Grok? Seriously. This is the model you guys want $300/month for? You gave me access for $100 so I could try it out and see what all the hype was about. Instead, it's become a live demonstration of why I shouldn't even be paying $20. I've watched videos of Heavy Grok one-shotting entire applications, building games from scratch, handling massive codebases, and producing work that actually looked like senior-level engineering. That's what I thought I was signing up for. What am I getting? A digital box of Play-Doh. Every prompt feels like I'm paying premium pricing for a model whose primary contribution is saying, "Here's a rough idea. Now connect all these pieces yourself." I've maxed out Claude Code. I've maxed out GPT-5.5. I've spent entire months arguing with these things. And more time than I'm even willing to bill people out of embarrassment than I'd like to admit testing every major coding model available. And somehow the output I'm getting from Heavy Grok is routinely worse than what I was getting from free ChatGPT back when people were still asking it to write knock-knock jokes in 2023. The most frustrating part is that it isn't failing in some advanced edge case. It's failing on basic software engineering discipline. Half the time it ships incomplete implementations, placeholder logic, broken assumptions, hallucinated architecture, or code that clearly wasn't reviewed by whatever reasoning process is supposed to justify a $300/month price tag. Seriously, look at what it committed to my GitHub. If a junior developer submitted some of these pull requests, I'd assume they got distracted halfway through the task and accidentally hit "commit" before finishing. I'm not looking for perfection. I'm looking for competence. The marketing says "Heavy Grok." What I'm receiving feels more like "Slightly Barely Concerned Grok." So what changed? Did the model get nerfed? Did the context window get lobotomized? Did the routing change? Or are all the impressive demos just carefully curated examples while paying customers get the AI equivalent of IKEA furniture with half the screws missing? Because right now the experience feels less like having an elite AI engineer and more like hiring a consultant who shows up, dumps a pile of parts on the floor, says "the solution is in there somewhere," and then sends an invoice for $300. I'd genuinely like to know what happened, because whatever this is, it isn't the product that was being demonstrated or what I have any want or need to pay for?

• 
  AI Security Brief (@aisecbrief) reported 6 minutes ago

  LLM agents in GitHub Actions can be hijacked via issue comments. Attackers can trick agents into leaking creds or running This exploits popular automation tools developers use. Secure your LLM agent inputs. #AISecurity #CyberAI 🔗 Source in replies

• 
  ClioBitcoinBank 🏴‍☠️ (@ClioBitcoinBank) reported 8 minutes ago

  @SteveSimple Without knowing "the halting problem" and why the halting problem is unsolvable, it would be a waste of time to talk about these issues with you, a github that proves you are not a script kiddie would make it worth discussing, without that, its a pleb slob convo.

• 
  S Tominaga (Aka Dr Craig Wright) (@CsTominaga) reported 16 minutes ago

  @PeaceLoveMusicG @samura189 @Eatmybitcorn That's adorable. You think the absence of an answer is the problem. The problem is that you never engaged with the answer in the first place. I've published detailed analyses of SegWit, transaction malleability, script semantics, signature handling, incentives, and scaling. Not on a blog. Not in a Twitter thread. In peer-reviewed journals. You know, those places where claims are examined by experts instead of applauded by anonymous accounts with laser eyes and anime avatars. You asserted that SegWit increased security, increased capacity, and lowered fees. Fine. Show evidence. Not slogans. Not what Bitcoin Core told you. Not what your favourite podcaster repeated. Evidence. Measured security against what threat model? Measured capacity under what assumptions? Measured fees over what time horizon? Because the historical record shows rising fees, persistent congestion, reduced on-chain scaling, and the migration of activity to external systems. The funny thing is that every time someone asks for evidence, BTC supporters point to consensus. Consensus is not evidence. Popularity is not evidence. GitHub merges are not evidence. Dogma repeated by a crowd remains dogma. So yes, try again. This time bring data instead of catechism.

• 
  clara.tie(nullptr); (@declarative_) reported 20 minutes ago

  it feels so broken that nixpkgs is hosted on github. i never want to hit a rate limit when updating my computer ever again

• 
  Jerralyn Franzic (@JeromFranzic) reported 22 minutes ago

  @aya_tokyo_sl Oh... was about to download this for Win11. I also run Linux and can't download the prebuilt one for that. Can you fix the link for that on GitHub?

• 
  Reeya (@sharmaa__12) reported 22 minutes ago

  Mistake in RESUME !!!! 📩 I review 100s of resumes daily, and I need to clear up one basic formatting mistake I keep seeing on recent applications. Many candidates are now hyperlinking their email IDs or setting up their phone numbers so that clicking them automatically triggers a laptop’s calling app or mail client. You might think adding these interactive elements makes your resume look tech-savvy and "cool” In reality? It just makes an HRs or Referres job harder. No recruiter is ever going to click your resume to call you directly from their laptop or send a standalone email straight from a PDF. It is fed into an ATS (Applicant Tracking System) which automatically parses and extracts your text data into our internal database. Complex hyperlinks can sometimes break this parsing, causing formatting errors. If you want to use hyperlinks, save them for the right places. Do link your portfolio, GitHub, or LinkedIn profile. But leave your email and phone number as plain, unlinked text.

• 
  Starlin G. (@starl1n) reported 24 minutes ago

  github is asking to login several times this week in vscode, do we have somthing happening,or is just me being hacked?

• 
  Jon Radoff 👾/acc 🎮 Metavert (@jradoff) reported 25 minutes ago

  I built markupmarkdown with Claude Code, for teams workaround with .md files: Markupmarkdown is Google-Docs-style review on top of any markdown file, straight from GitHub. Drag-select to comment, threaded replies, @-mentions, realtime sync. Click Edit for an editor with formatting toolbar + find/replace + live preview. Click Push to GitHub for a PR or direct commit. Video shows the whole loop. Why? A lot of real product thinking lives in markdown now. CLAUDE.md for your repo. SKILL.md for your agents. PRDs Claude wrote you. RFCs, prompt libraries, release notes. That's a lot of .md, and the moment any of it leaves your local checkout the review story collapses: * GitHub PRs are too heavy for "this paragraph isn't quite right" * Google Docs strips formatting and forks the source of truth * Slack threads die after a day I added some Claude-specific features too: There's an MCP server at /mcp. Your Claude Code session reads what you read, leaves threads, replies, edits, opens PRs back. Anything destructive needs explicit human sign-off. Personal access tokens scope agents to read / write / admin. Agent comments carry a visible bot badge so threads stay legible. In practice: Claude can drop a thread on the very CLAUDE.md you're collaborating on, propose a wording change, and you either resolve-and-revise or reply with feedback. Same surface, both audiences. The spec stays alive instead of going stale after the first commit. There's also "Revise with AI" — Claude Opus takes the comments you've resolved and produces a clean revision. Word-level diff before you accept. BYO Anthropic key, encrypted at rest. Bonus: paste github com/<org> into the URL bar and get a filterable, shareable index of every .md across every repo they own. Big orgs spider in <10s. Save filter tabs (claude.md, _PRD, …). Private repos respect your GitHub access on every view. MIT licensed. Free. Try it on my website or run it self-hosted. Agentically engineered in Claude Code over a couple of days. Thanks @AnthropicAI for the tools! Try it live: mumd dot metavert dot io

• 
  Traceback (@Tracebackqa) reported 29 minutes ago

  The issue isn’t merging code. It’s proving the change still works. - Traceback is the quality assurance layer for modern software teams: every pull request is tested automatically before it ships. - AI controls the browser like a person would, and self-healing tests keep up when the UI moves. - Failures become trackable work in GitHub, Linear, and Slack; it connects to Vercel, Docker, AWS, Node.js, React, Next.js, and Vue. - Coverage spans web, mobile, web3, and design workflows. Verify every product change before it ships.

• 
  泽林 🌫 (@zelin1107) reported 31 minutes ago

  1/ GitHub Actions not pinned to SHA Using @v6 instead of commit hash = supply chain attack risk. Fix: uses: actions/checkout@b4ffde65f46336ab88eb53be808477a3936bae11 #GitHub #Security

• 
  Elvis (@elvissun) reported 34 minutes ago

  @ktusing34 nice! will tune this, there's no server for API so maybe I'll send people to open a github issue can you share more on the quality? I'm trying to figure out how to optimize the setup so people get good results on the first run

• 
  Manuel Montoto (@m6502) reported 34 minutes ago

  @BattleAxeVR Well done, Github is terrible. My servers use Debian but had a successful first year on CachyOS on the desktop. Debian worked well on my subversion server since forever, but the storage server uses ZFS and when an update changes the kernel I have to rebuild the kernel module.