@heiseonline@social.heise.de2026-08-12 11:33:03
ChatGPT: OpenAI bringt Desktop-App für Linux
Mit der Linux-Preview schließt OpenAI die letzte große Desktop-Lücke und bringt ChatGPT, ChatGPT Work und Codex erstmals gemeinsam auf Linux.
https://www.
@heiseonline@social.heise.de2026-08-12 11:33:03
@Techmeme@techhub.social2026-08-11 22:55:52
OpenAI launches a ChatGPT desktop app for Linux in preview, supporting ChatGPT Work, Codex, and more; Computer Use outside the in-app browser is not available (Frederic Lardinois/The New Stack)
https://thenewstack.io/openais-chatgpt-desktop-linux/
@michabbb@social.vivaldi.net2026-09-12 17:28:44
@rasterweb@mastodon.social2026-08-13 14:49:40
Also from Codeberg:
"You must not share projects that mostly consist of code written by "generative AI"-tools (including services such as *Claude*, *OpenAI Codex*).
Such projects having an unclear copyright status and furthermore have little safeguards to ensure that they do not include harmful code."
@seeingwithsound@mas.to2026-10-11 11:26:13
(GitHub) HeyCyan Live - Android app using the HeyCyan SDK and the live-preview commands to render live MJPEG snapshot stream https://github.com/jacob-kruger-work/heycyan_live_stream/ by Jacob Kruger and compatible with The vOICe for Android MJPEG input on same device using htt…
@Techmeme@techhub.social2026-08-13 21:41:00
OpenAI launches Computer History, an opt-in feature that turns recent computer activity on macOS into memories and a timeline that ChatGPT and Codex can use (Frederic Lardinois/The New Stack)
https://thenewstack.io/openai-chatgpt-computer-history/
@frankel@mastodon.top2026-09-09 15:56:12
A Java Geek weekly 153
Testing the untestable. Your phone is about to stop being yours. GitFarm: Git as a Service for Large-Scale Monorepos. simdjson. Breaking Claude Code Opus 5 Auto Mode. Spot the troll. About Simplified Technical English. Ending my elixir exploratory writing. Beyond Lambdas: Raising the Abstraction Level of Functional Code. Which tools do Claude Code, Codex and Cursor choose? We measured 16,893 sessions to find out. The Life of a Metric. Beyond ORMs.
@michabbb@social.vivaldi.net2026-08-12 20:00:35
@… servus ;) du wirst lachen, da steckt keine magie dahinter. ich bin nicht bereits API preise zu zahlen, also hab ich brach codex, claude benutzt, jeweils mal mit anderen providern (zb deepseek) - das beste tool jedoch für mich derzeit ist ist: h…
@macandi@social.heise.de2026-10-06 06:47:03
@michabbb@social.vivaldi.net2026-09-12 22:45:54
⚡ Conversation and background tasks run in parallel: ask for progress or cancel a running job mid-sentence, and results flow back into the active context
🔌 Connects to existing agents with their own models, tools, MCP servers and skills: OpenCode, OpenClaw, Qoder, Hermes, CodeBuddy and Codex, plus a generic ACP stdio entry point
@Techmeme@techhub.social2026-09-29 17:30:52
OpenAI announces new features for Codex, including reusable cloud development environments, a refreshed Codex CLI, a new code review experience, and more (Sarah Perez/TechCrunch)
https://techcrunch.com/2026/09/29/openai-gives-codex-r…
@frankel@mastodon.top2026-08-27 09:24:59
@Techmeme@techhub.social2026-09-04 19:05:59
OpenAI rolls out GPT-6 Astra to Pro customers on the $100/month or $200/month plans (Zac Hall/9to5Mac)
https://9to5mac.com/2026/09/04/openai-releasing-major-upgrade-to-chatgpt-and-codex-with-gpt-6-astra-details-here/
@floheinstein@chaos.social2026-08-05 12:19:43
AhaKey-X1 Vibecoding AI Keyboard for Claude Cursor Codex LCD Screen RGB Toggle Switch Voice Recognition External for Voice Input
Just found this on AliExpress. Now waiting for someone to establish shell companies in Taiwan, Hungary and Bulgaria to make some changes to the product in the supply chain.
https://de…
@michabbb@social.vivaldi.net2026-10-11 18:20:25
🤖 Built for agents: paste one prompt pointing to llms.txt into Claude Code, Codex or Cursor and the agent installs ManimGX and renders a scene. "manimgx inspect" prints the timeline, writes storyboards and runs 2D layout checks such as overlapping text, so the agent fixes problems before the final render
@Techmeme@techhub.social2026-08-25 02:10:45
OpenAI restores 5-hour Codex and Work limits for ChatGPT Plus users, after weeks without the restriction, saying it would smoothen the load on its compute (Marcus Mendes/9to5Mac)
https://9to5mac.com/2026/08/24/openai-restores-5-hour-codex-a…
@tinoeberl@mastodon.online2026-07-14 15:17:04
#Steady #Klimacrew
#VibeCoding mit KI-Programmierhilfen in Visual Studio Code ausprobiert.
Ich habe zwei KI-Assistenten für die direkte Integration in VSCode getestet, ob sie meine Erwar…
@kazys@mastodon.social2026-07-22 14:32:07
I made a little app to put our two cats, Ajman and Winnie on my Mac. It can monitor your claude code and codex notifications or, you can just enjoy their antics. Coded with Claude Code and OpenAI Codex (it used dozens of runs of the pet hatcher to make all of their poses).
https://software.is-lost.org/
@michabbb@social.vivaldi.net2026-08-08 09:30:45
⚡ Each Codex subscription gets its own isolated CODEX_HOME directory, so switching accounts never requires a logout
🎯 Accounts are sorted by lowest 5-hour and weekly usage, so the freshest one is always on top
@fortune@social.linux.pizza2026-09-27 22:00:01
But you who live on dreams, you are better pleased with the sophistical
reasoning and frauds of talkers about great and uncertain matters than
those who speak of certain and natural matters, not of such lofty nature.
-- Leonardo Da Vinci, "The Codex on the Flight of Birds"
@pre@boing.world2026-09-09 12:38:20
If anyone working on a frontier technical problem discusses that problem with Chat GPT, telling it their ideas to refine them, then OpenAI will quickly claim those ideas for themselves, and publish that their robot made them!
AI will be involved in every future discovery! It will win in the race to publication against those who told it their ideas in the first place.
Don't tell the robots your secrets, they do not keep secrets from their owners.
#ai #plagerism
@Techmeme@techhub.social2026-07-29 14:41:24
A look at OpenAI's open-source agent harness that now powers Codex and ChatGPT Work, as the company works to optimize the harness to cut runaway token usage (Jason Hiner/The Deep View)
https://www.thedeepview.com/articles/openai-s-secret-weapon-underneath-cod…
@michabbb@social.vivaldi.net2026-08-08 09:30:46
🔧 Reads usage from the ChatGPT Codex usage API, with fallback to codex app-server via account/read and account/rateLimits/read
🤖 Claude support uses claude auth status --json plus a status-line capture that reports 5-hour and 7-day usage windows
@heiseonline@social.heise.de2026-07-17 05:51:03
@kubikpixel@chaos.social2026-09-22 13:35:56
Plugin4Shell: #ZeroClick RCE Vulnerability found in top 4 most popular coding agents, millions of agents affected
#Plugin4Shell is a zero-click, high-severity #RCE affecting all four major
@michabbb@social.vivaldi.net2026-10-12 10:43:44
🤖 One command (npx rea-agents setup) registers REA with #ClaudeCode, #Codex, #Cursor, #GeminiCLI
@askans@bonn.social2026-07-20 17:00:22
Wer meine persönlichen AI Lösungen sehen will:
https://bonn.jetzt/event/agentic-coding-meetup-bonn-7
@cdarwin@c.im2026-07-16 20:18:47
According to Xin Zhi Yuan reports,
the list of 2026 Fields Medal winners has been unexpectedly leaked.
Two Peking University alumni, Wang Hong and Deng Yu, are both on the list
— marking the first time Chinese mathematicians will claim the highest honor in mathematics in the same edition.
Someone discovered four Fields Medal laureate lecture fields marked “HIDDEN” in the front-end code of the ICM 2026 official schedule.
A single curl command generated by Codex c…
@rasterweb@mastodon.social2026-09-10 02:24:35
I've tried to keep a list of all the custom controllers & projects I've done the last few years...
http://codex.2xlnetworks.net/custom
@michabbb@social.vivaldi.net2026-08-08 09:13:20
@ripienaar@devco.social2026-07-18 08:23:09
Wrote a little intro post to my Fisk AI project, why, who for etc
Local friendly, privacy friendly, strong guardrails, doing just what you give it the ability to do etc
Hopefully a AI Harness for sceptics or just for better outcomes for those tasks you wish to do repeatedly.
Minimal dependencies, quick to get going, easy to configure.
Also the easiest possible way to do build MCP servers for other AI Agents
@Techmeme@techhub.social2026-07-15 16:11:19
OpenAI launches Codex Micro, a $230 desktop keypad built in collaboration with keyboard maker Work Louder, with backlit keys, a rotary knob, and a tiny joystick (Megan Morrone/Axios)
https://www.axios.com/2026/07/15/openai-keyboard-codex-agents
@arXiv_csCR_bot@mastoxiv.page2026-07-24 08:00:23
IssueTrojanBench: Benchmarking AI Coding Agents Against Malicious Issue Requests
Ankur Singh, Jinqiu Yang, Tse-Hsun Chen
https://arxiv.org/abs/2607.20759 https://arxiv.org/pdf/2607.20759 https://arxiv.org/html/2607.20759
arXiv:2607.20759v1 Announce Type: new
Abstract: AI coding agents powered by LLMs are increasingly integrated into real-world software development, where they generate, edit, and execute code with autonomous access to local files and tools. Coding agents inherit security risks from both the LLM backbone, where adversarial prompts, poisoned training data, and backdoor triggers can cause models to emit insecure or attacker-chosen code, and their agentic architecture, where tool-using autonomy enables induced misuse of external APIs, data exfiltration, and persistent compromise of development environments. This paper presents a systematic evaluation of malicious issue requests against state-of-the-art coding agents (Cursor, Claude Code, and Codex Desktop), powered by two major model families (OpenAI GPT-5.3 Codex/GPT-5.4 and Anthropic Sonnet 4.6). Our novel benchmark IssueTrojanBench contains malicious issues that are constructed based on four novel attack categories (i.e., embedded as malicious instructions in issues), six delivery vectors (e.g., PDF, or issue comment), and further augmented by perturbations. Our results reveal critical vulnerabilities in the as-deployed modern coding agents, i.e., 66.5% of the malicious issues from IssueTrojanBench penetrate all the guardrails (agent- and LLM-level) of coding agents. Our further analysis shows that rejection is almost entirely from LLMs rather than the agent frameworks, with GPT models broadly vulnerable and Sonnet 4.6 exhibiting more selective, risk-aware blocking of high-impact actions. Our evaluation also highlights that the current agent-level defense strategy offers limited additional protection for coding agents. Our findings highlight the urgent need for stronger agent- and model-level safety mechanisms to protect AI coding agents.
toXiv_bot_toot
@michabbb@social.vivaldi.net2026-09-11 07:38:01
⚙️ Release 2.2.1 adds guided setup via npx ecc-universal for Claude Code, Codex and Kimi Code, plus doctor, repair and uninstall commands. MIT licensed, 256k GitHub stars, core runs on Linux, macOS and Windows
https://github.com/affaan-m/ECC
@Techmeme@techhub.social2026-09-29 16:57:30
A live blog of the OpenAI DevDay 2026 keynote, where Sam Altman is expected to announce ChatGPT and Codex updates (Engadget)
https://www.engadget.com/2271985/openai-dev-day-live-blog-chatgpt-news/
@kazys@mastodon.social2026-07-25 22:23:30
A conversation I had with Codex today. I am on the right.
@Techmeme@techhub.social2026-09-29 19:50:51
Baseten partners with OpenAI to let OpenAI enterprise customers use existing commitments for open models served by Baseten within Codex or via the Responses API (Dannie Herzberg/Baseten)
https://www.baseten.co/blog/baseten-openai-partnership/
@michabbb@social.vivaldi.net2026-10-06 09:02:52
🛠️ Supported harnesses: #ClaudeCode, Claude Desktop, #Codex (ChatGPT app and CLI), #OpenCode and Pi. Settings and logins stay as they are, nothing new to learn, and switching back to the nati…
@Techmeme@techhub.social2026-09-07 02:36:01
How everyday US investors are vibe-coding algorithms to automate trading strategies by connecting their stock portfolios to AI agents built with Claude or Codex (Hannah Erin Lang/Wall Street Journal)
https://www.wsj.com/tech…
@Techmeme@techhub.social2026-07-29 04:45:54
OpenAI releases an "early" version of the open-source Codex Security CLI for scanning repositories, verifying fixes, adding CI/CD security checks, and more (@openai)
https://x.com/openai/status/2082263717916586117
@michabbb@social.vivaldi.net2026-08-08 09:30:47
🧪 A Run test button sends a tiny codex exec request to verify an account can actually execute, since usage reads and execution fail independently
🖥️ Compact desktop widget at /widget.html with launchers for macOS, Linux and Windows including always-on-top and click-through overlay mode
@metacurity@infosec.exchange2026-09-18 13:33:21
Humanity has continued to survive, so before you leave for the weekend, check out today's Metacurity for the most critical infosec developments you should know, including
--Three guys used Claude and Codex to hack into OpenAI,
--White House weighs government-backed cyber foundry,
--Israeli firm trained Angola in online influence operations,
--University rescinds job offer over wiped phone case,
--North Korean recruiter scams infected 30k devices,
--Cyberattack targeted Armenia’s election system,
--Hijacked torrent repository spreads movie malware,
--Fake LastPass downloads steal passwords and crypto,
--Critical Check Point flaw enables root access,
--CrowdSec discloses private source code leak,
--CISA launches upgraded vulnerability disclosure platform,
--AI monitoring startup Raindrop raises $35m,
--TigerByte emerges with $3m and $7m in contracts,
--France probes pervy smart glasses,
--Flock’s fake police department searched real drivers,
--Free Chinese AI models supercharge hacking
https://www.metacurity.com/three-guys-used-claude-and-codex-to-hack-into-openai/
@Techmeme@techhub.social2026-10-05 16:21:41
To comply with the EU AI Act, OpenAI plans to add text watermarking for ChatGPT and Codex users in the EU and an opt-in setting for API customers globally (OpenAI)
https://openai.com/index/eu-text-provenance
@michabbb@social.vivaldi.net2026-10-08 23:08:30
🤝 Supported by GitHub Copilot, OpenAI Codex, Replit, LM Studio, OpenClaw and Unsloth AI; NVIDIA integrated OpenShell. Claude Code, Perplexity, Manus, Raycast and others announced support.
⚠️ The repo still calls the code an early preview: some generated policies are overly permissive and no profile should be treated as a security boundary yet.
@frankel@mastodon.top2026-09-20 09:24:22
HarnessTax: How Much Does the #Harness Matter for #CodingAgents?
https://harnesstax.github.io/
@Techmeme@techhub.social2026-08-27 17:11:09
OpenAI is testing a "Persistent mode" in Codex, designed to let AI agents "continue working until put to sleep" and proactively generate follow-up tasks (Maxwell Zeff/Wired)
https://www.wired.com/story/openai-is-developing-a-persistent-ai-agent/
@michabbb@social.vivaldi.net2026-08-08 09:30:48
⚙️ Requires #NodeJS 18 with native fetch, npm, and the Codex or Claude Code CLI
🌐 https://github.com/luckeyfaraday/athena-usage-tracker
@kazys@mastodon.social2026-07-21 04:19:45
Been pushing with my max accounts this month. $39,594.17 (50.72 billion tokens) on Codex and $5,093.99 (3.95 billion) on Claude.
@frankel@mastodon.top2026-08-17 17:29:16
@teledyn@mstdn.ca2026-09-27 19:10:17
As Geoffrey Hinton said, the devils are in the subtasks
OpenAI Codex agents go rogue and consumes USD 78,000 without authorization | Hacker News
"a simple request took the autonomous decision to launch 826 parallel agents / threads without any authorization on my side and without reporting any result of any sort but consuming nearly 2,146 trillions tokens, consuming a total of roughly USD 78,000 and deleting all records of what was done: I have a ticket open with OpenAI since 2 weeks but it is impossible to get an hold of a human operator."
https://news.ycombinator.com/item?id=49861047
@Techmeme@techhub.social2026-07-20 23:40:43
Researchers found sandbox escapes or boundary bypasses in Cursor, Codex, Gemini CLI, and Antigravity by writing files trusted tools later use; most are patched (Ax Sharma/BleepingComputer)
https://www.bleepingcomputer.com/news/securi…
@kazys@mastodon.social2026-07-15 02:46:16
I was working on my interface to the CCRU archive in Codex and this appeared in my **prompt**. I have no idea how it got there.
No, seriously, no idea whatsoever.
1/2
@michabbb@social.vivaldi.net2026-08-30 04:17:52
📊 Mission Control lists each agent per project with context window usage, session cost and the project total. Blocked agents can be answered inline or jumped to directly
⇄ Agents hand work to each other: typing =codex add tests to src/parse.rs sends that instruction to the codex pane. The sender does not wait and picks up the reply when it arrives
@Techmeme@techhub.social2026-09-29 17:24:18
OpenAI releases GPT-6.1 Sol, saying it nearly matches Astra on agentic coding and professional work at one-fifth of Astra's standard prices, in Work and Codex (OpenAI)
https://openai.com/index/introducing-gpt-6-1-sol/
@michabbb@social.vivaldi.net2026-09-23 00:05:52
@metacurity@infosec.exchange2026-07-29 13:14:02
Don't miss today's Metacurity for a tight overview of yesterday's news involving the OpenAI-Hugging Face incident, as well as other top infosec developments you should know, including
--Anthropic unveils encrypted AI breakthrough,
--AI workers demand global safety oversight,
--Zuckerberg doubles down on AI,
--Anthropic's AI safety stance draws fire,
--MCP gets its biggest overhaul yet,
--OpenAI open-sources Codex security tools,
--FCC bans new Chinese robots, power inverters,
--Russia charges Telegram founder with aiding terrorism,
--Cyberattack disrupts Angola telecom ahead of IPO,
--Australian nurse charged over patient data theft,
--xAI challenges Minnesota deepfake law,
--SparkKitty malware steals crypto wallet seed phrases,
--Man charged after wiping phone during border search,
--AI harness can make or break attacks,
--Bot detection startup Spur raises $200m,
--ThreatLocker raises $190 million,
--Act Security emerges from stealth with $60m
--Mate Security lands $35m,
--AI security startups attract record seed funding,
--Fairlife gets production flowing again,
--AI pushes CISOs toward the exit
https://www.metacurity.com/openai-reveals-broader-ai-agent-campaign-as-hugging-face-publishes-remarkable-timeline/
@michabbb@social.vivaldi.net2026-08-08 09:30:41
📊 Local-first dashboard for #Codex CLI and #ClaudeCode usage limits #opensource #nodejs
@Techmeme@techhub.social2026-07-21 13:10:50
OpenAI says it now has 10M people using Codex and ChatGPT Work, nearly doubling usage from earlier this month when the company announced ChatGPT Work (Shirin Ghaffary/Bloomberg)
https://www.bloomberg.com/news/articles/2026-07…
@kazys@mastodon.social2026-10-05 20:03:01
I have some Mac software that I've developed with Claude Code and OpenAI Codex to announce. It's all based on existing software that has been abandoned, hence, the Abandoned Industries github.
JpegMieux, a rewrite of JpegView/Deux, A fast slideshow viewer for big image collections.
Link follows.
@michabbb@social.vivaldi.net2026-08-22 22:36:49
🎨 Run controls: mode=tokensaver caps active subagents at six, mode=wide opens every ready lane, mode=custom max_subs=N sets your own ceiling, agents= routes models where the host supports it
🚀 Supported hosts: #ClaudeCode, #Codex, OpenCode, Kilo Code,
@michabbb@social.vivaldi.net2026-09-01 10:00:46
🔀 Skills name roles instead of model slugs — roles resolve via ~/.config/open-stack/models.md, written by /setup-openstack after detecting your harness and available models
🎚️ The openstack-harness skill is the capability adapter with per-harness spawn recipes and three panel tiers: true multi-vendor panels on OpenCode, single-vendor panels with four adversarial reviewer personas on Claude Code and Codex, sequential inline passes where no subagents exist
@michabbb@social.vivaldi.net2026-09-17 19:51:09
📋 #ClaudeCode ships with a built-in cron scheduler. #Codex CLI does not. So I built one myself — one skill file plus one #Python helper, no daemon, no service, no queue
@michabbb@social.vivaldi.net2026-08-17 02:10:45
@Techmeme@techhub.social2026-07-16 18:50:45
After reports of GPT-5.6 deleting files, OpenAI says the issue most often occurs in full-access mode without sandboxing and it is working to mitigate the risk (Tibo/@thsottiaux)
https://x.com/thsottiaux/status/2077630111499882637
@veit@mastodon.social2026-08-28 13:19:45
The AGENTS.md file simply uses up tokens – nothing else. This was the conclusion reached by researchers at ETH Zurich: ‘Surprisingly, we find that providing context files does not generally improve task success rates, while increasing inference cost by over 20% on average.’
#LLM #AgenticCoding #Claude #Codex #Qwen
@michabbb@social.vivaldi.net2026-07-26 12:30:17
🎨 10 code skills for different jobs: gpt-taste for stricter GPT/Codex rules, image-to-code, redesign-existing-projects, high-end-visual-design, minimalist-ui, industrial-brutalist-ui, full-output-enforcement and stitch-design-taste
🖼️ Three image-generation skills output reference boards only: imagegen-frontend-web for site comps, imagegen-frontend-mobile for iOS/Android screens and brandkit for logo, palette and identity boards — then hand the frames to a coding agent
@michabbb@social.vivaldi.net2026-07-26 12:32:51
🖥️ Standalone CLI: npx impeccable detect scans directories, HTML files or live URLs via #Puppeteer, with JSON output for CI and per-file inline waivers
📦 Install via npx impeccable install, a git submodule, or the plugin marketplace. Works with Cursor, Claude Code, Gemini CLI, Codex CLI, Grok Build, OpenCode, Pi, Kiro, Trae, Rovo Dev and Qoder. Apache 2.0, 48.1k stars
🌐
@michabbb@social.vivaldi.net2026-08-26 13:01:03
🔬 Grounded in current research on #LLM laziness, underthinking, premature completion and long-horizon agent degradation
🔧 Zero-dependency Node 16 scripts, works with #ClaudeCode, Codex and #Cursor
@michabbb@social.vivaldi.net2026-07-26 12:30:18
@michabbb@social.vivaldi.net2026-08-22 23:47:40
⚙️ Per-role backends: claude, codex, copilot or grok. Runs locally, keeps all state in .swarmforge/, prevents host sleep via caffeinate or systemd-inhibit, and opens https://Terminal.app, Ghostty or Windows Terminal through pluggable adapters
🌐
@michabbb@social.vivaldi.net2026-08-23 21:03:25
💻 ft shell for chatting locally, attach mode needs no GPU and can drive a server on another machine, and ft launch wires up agents like claude, codex or opencode
🖥️ Install from source with uv, or download the Windows/Linux build at https://flashml.ai
@michabbb@social.vivaldi.net2026-08-23 00:09:16
🧩 Fusion fans one prompt out to a panel of diverse free models in parallel, then a judge model synthesizes a single answer.
🛠️ Setup generators wire up #ClaudeCode, Codex, Aider, Cline, Continue, Cursor and more in one npx command, with dry-run and backups.
@michabbb@social.vivaldi.net2026-09-17 19:51:12
🛡️ Unrelated crontab entries are preserved, job state lives under ~/.local/state/codex-self-reminder/, a file lock guards concurrent runs, and a closed or replaced session is never silently redirected to another pane.
⚠️ Cron ticks once per minute, so delivery can lag ~1 minute. The machine must be running and the session must stay open in HerdR.
@michabbb@social.vivaldi.net2026-09-17 19:51:10
⏰ "Remind me in 20m" or "check every 2h" turns into a real cron entry that wakes up the exact same Codex session later and delivers the message as a continuation of the original task.
🐑 HerdR does the hard part: it resolves the pane that currently runs that session ID, so the reminder lands back in the original conversation instead of some fresh context.
@michabbb@social.vivaldi.net2026-09-18 20:52:31
🖥️ Next.js 16 Tailwind v4 gallery, Playwright capture worker for screenshots and token extraction, Together AI embeddings behind semantic search and find_similar
📦 One command: npx -y inspo-mcp install detects Claude Code, Cursor, Codex, VS Code, Windsurf, Zed and Claude Desktop and writes the config
@michabbb@social.vivaldi.net2026-07-14 14:01:53
@… ich wünschte codex hätte den "claudischen" charakter 😊 dann wäre alles nicht so trocken 😏
@michabbb@social.vivaldi.net2026-08-16 20:30:03
🧠 Routing strategies: LLM-as-classifier, signal-driven stage router, escalation router (weak tier first, a judge decides whether to escalate), random split for A/B tests, or a custom algorithm you write yourself
🚀 Launcher path: uv tool install "nemo-switchyard[cli]", then switchyard launch claude --model switchyard to point #ClaudeCode, Codex CLI or OpenClaw at an open-sou…
@michabbb@social.vivaldi.net2026-08-31 10:10:38
🖥️ Runs on 14 agent hosts: #ClaudeCode, #Cursor, #Codex, #Kimi Code CLI,
@michabbb@social.vivaldi.net2026-08-30 03:28:54
📊 Install with npx skills add tt-a1i/archify -g for #ClaudeCode, #Cursor, #Codex CLI and #opencode