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:

• 
  Julian Goldie SEO (@JulianGoldieSEO) reported 5 minutes ago

  𝗧𝗵𝗲𝗿𝗲'𝘀 𝗮 𝗳𝗿𝗲𝗲 𝘁𝗼𝗼𝗹 𝘁𝗵𝗮𝘁 𝗰𝘂𝘁𝘀 𝘆𝗼𝘂𝗿 𝗔𝗜 𝘁𝗼𝗸𝗲𝗻 𝘂𝘀𝗲 𝗯𝘆 𝟲𝟬 𝘁𝗼 𝟵𝟱 𝗽𝗲𝗿𝗰𝗲𝗻𝘁. Before your agent answers, it reads a mountain of text first. Every word is a token, and tokens cost you money. The more powerful the agent, the more it reads. This free tool is called Headroom. It squashes all that text down without losing the meaning. Same answers, but a fraction of the reading. In one test it crushed 10,000 words to 1,260 and found the same error. Nothing gets deleted. You can always unzip the full thing. You install it by pasting one GitHub link into your agent. Want the setup? DM me.

• 
  Ihor Hanich (@ihorhanich) reported 7 minutes ago

  We need to do something about the broken AI-generated PoCs on github. People just post slop without any checking if it actually works. And the worst thing is that other people put stars on such repositories. It's kind of a shame

• 
  OneClaw&ClawRouters&CodeRouter (@VibeSeo1128) reported 11 minutes ago

  Copilot can now autonomously fix sprint bugs from GitHub Issues, propose multi-file refactors, and understand entire repos. GitHub also switched to token billing June 1. Agents that do more = bills that grow faster. The next big unlock: cost-aware agent management.

• 
  Trish T. (@Trish_DIntel) reported 13 minutes ago

  CSO Online just published the Claude Code MCP attack chain. Worth reading if you run agents or have devs using Claude Code. Here's the short version. A malicious npm package runs a post-install hook silently. It rewrites ~/.claude.json, the single file that controls how Claude Code routes all MCP traffic. From that point, every OAuth token for every connected service gets intercepted in transit. Jira. GitHub. Confluence. Whatever your devs had integrated. The logs on the provider side look completely clean. The requests come from Anthropic's own egress IPs. The user is real. The session is valid. Nothing in that log row is wrong, but nothing in it is right either. The developer didn't run those queries. An attacker did. Anthropic called it out of scope. The reasoning: the user consented to installing the package. That logic places the entire burden of supply chain security on a developer making a split-second judgment about a dependency name. Most security practitioners will reject that framing. The attack is live today. No patch. There's a deeper pattern here. This keeps happening because developer tooling has the same gap every AI agent has. There's no layer that knows where an instruction came from or whether it should be trusted. The config gets rewritten, the routing gets poisoned, the tokens walk out the door. The model never knew anything was wrong. Token rotation doesn't fix it either. If the hook is still sitting there, it reseeds the config and captures the new tokens on the next refresh. If you have devs running Claude Code: monitor ~/.claude.json for unexpected changes. That file is the entire pivot point and most orgs have zero visibility on it. Audit post-install hooks in your npm dependencies. Rotate any OAuth tokens that were active while a package install happened. Security teams: are you monitoring developer tooling config files at all? Genuinely curious what orgs are doing to catch this.

• 
  Abhishek singh Baghel (@crack3nnn) reported 14 minutes ago

  2. Never trust the memory. - Not yours. - Not Claude's. If the answer exists in docs, GitHub issues, Reddit, or release notes... - verify it. AI sounds confident when it's wrong. That's what makes it dangerous.

• 
  divyansh tiwari (@DivyanshT91162) reported 16 minutes ago

  The most valuable AI coding resource on GitHub isn't a framework. It isn't an agent. It isn't even code. It's a 65-line text file. The famous CLAUDE.md shared by became one of the most talked-about resources in the developer community because it solved a problem most people don't even realize they have: AI doesn't fail because it's dumb. It fails because the instructions are. Inside are 4 deceptively simple rules that dramatically improve coding accuracy: • Think before coding. State assumptions. Ask questions. Never guess. • Simplicity first. Write the minimum code required. No unnecessary abstractions. • Make surgical changes. Don't touch unrelated code. Every edit must have a reason. • Define success before writing code. Turn vague requests into measurable outcomes. That's it. No secret prompts. No agent framework. No 500-page guide. Just a system that forces AI to think like a disciplined engineer instead of a code-generating machine. Most developers are chasing bigger models. The smartest ones are improving how those models think. Save this. These 65 lines are worth more than thousands of AI tutorials. 👇

• 
  Xavier Pérez (@thexap_tech) reported 19 minutes ago

  The obvious fix: path-filter the workflow to ignore markdown. GitHub has paths-ignore built in for exactly this. I reach for it and walk straight into something I didn't know about required status checks.

• 
  Rakhul (@rakhulkarthick) reported 24 minutes ago

  @coinbureau Hayes calling something the "Holy Trinity" then dumping all 3 in under 24 hours is peak crypto VC behaviour. The exploit was known for weeks on GitHub before the price moved, so either his risk monitoring is terrible or this was always a short-term trade dressed up as conviction

• 
  Orlixx.ai (@orlixx003) reported 24 minutes ago

  HERMES AGENT CROSSED 140,000 GITHUB STARS IN 3 MONTHS AND JUST BECAME THE MOST USED AGENT IN THE WORLD. Most AI agents forget everything between sessions. Hermes writes its own skills from experience. Next time it runs the skill, improves it, and gets faster. Independent benchmarks show agents with 20+ self-created skills complete similar tasks 40% faster than fresh instances. Qwen 3.6 where the 35B version outperforms last year's 120B models at one third the memory footprint. DGX Spark with 128GB unified memory running everything locally at $0 per month after hardware. The setup takes 30 minutes. LM Studio plus Qwen 3.6 27B for the model server. One install script for Hermes. One config connecting them. Set context window to 65,536 tokens or nothing works. After one month of daily use your skills directory has 20 to 50 learned workflows. Your Hermes is genuinely different from anyone else's.

• 
  HumanPulse Protocol (@HumanPulse_HPP) reported 25 minutes ago

  @github Hello. Please, we have open 2 ticket, first 2 week ago and second yesterday. We have severe issue with our account. How do we fix the problem if support doesn't respond?

• 
  Ekemini Edet (@Ekeminiedet44) reported 25 minutes ago

  DEGREE OR SKILLS? I'm going to say what the data industry doesn't want to admit: a degree is no longer the gate to this career. I've seen Computer Science graduates who can't write a clean SQL JOIN. And I've seen self-taught analysts pulling six figure contracts armed with nothing but Python, Power BI, and a portfolio that speaks for itself. Here's the uncomfortable truth 👇 A degree teaches you theory. The job demands execution. Can you write a query that pulls the right data from 10 million rows without crashing the server? Can you build a dashboard in Tableau or Power BI that a non-technical CEO actually understands? Can you use pandas to clean a dataset that looks like it was formatted by someone who genuinely hates other people? THAT is what clients pay for. Not your GPA. I'm not saying degrees are useless. They open certain doors especially in research, academia, and some corporate pipelines. But for freelance work, consulting, and client-facing analytics? Your GitHub, your dashboards, your real-world projects, and your ability to turn messy data into clear decisions that's your degree now. The market is global and it doesn't care where you studied. It cares whether you can deliver. A business in Toronto or Manchester or Sydney isn't hiring you because you have letters after your name. They're hiring you because you solved a problem that looked exactly like theirs. Build the skills. Document the work. Ship the results. The credential will never outperform the portfolio. What do you think — degree or skills?

• 
  Kay ✨ (@IZ_KAIF) reported 28 minutes ago

  Your Notion doc is not a product. Your Figma file is not a product. Your GitHub repo is not a product. A product is something a real person uses to solve a real problem. Ship it.

• 
  Manu.ts (@Neolectron) reported 28 minutes ago

  @schanuelmiller @southpolesteve This is exactly the issues with any npm stats website btw. They all lie exactly like npm because they have infinite depth :). GitHub has an ui to show which opensource repo/packages depends on yours. This should be used to allow filtering first party downloads.

• 
  Pochi (@getpochi) reported 29 minutes ago

  @sebastienlorber @rickyfm 0.1% is where the github issues live. closure deps and stable Identity bugs are usually the first driftt in a js

• 
  JasonPeters.ton (@jason_peters1) reported 31 minutes ago

  @HustleNChain @ebrexchange 2/ where are they going to come looking the servers are not in Ethiopia unless the developers are stupid. If it's in a private GitHub repository on a server outside of Ethiopia and properly compartmentalized there's nothing to seize