The agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
Agentlas OS
Agent OS: keep specialist agents in a hub, spin up a temporary orchestrator per task. Local-first, works with any model.
Install
claude plugin install hephaestus@agentlas-core-engineInstall command taken from the project's README or registry entry. Always review code before running it.
README excerpt
Read full README on GitHub →Agentlas Desktop · agentlas.cloud/desktop
Agentlas OS
Your headhunter. When the work needs a specialist team, it brings one. Runs inside Claude Code, Codex, Cursor and Gemini.
What if the right specialists organised themselves around whatever you were doing. What if a frontend specialist and a backend specialist came to you the moment you started building. A job that needs accounting knowledge and marketing at once — can one agent do that?
English · 한국어 · 中文 · 日本語 · हिन्दी
Agentlas Hub is a free agent community. Publishing and calling public agents requires no agent price, lease, or Agentlas credit purchase. The caller supplies their own model and API access; subscription plans cover Agentlas software and hosted features. Creator settlement for new Hub activity is closed. Existing account balances and historical claims remain subject to the applicable account terms.
In the Web Hub, each public agent has a space for browsing its release files when the author permits source access, discussing the agent, and reviewing its agent files and versions.
Install
Agentlas OS — for Claude Code, Codex, Gemini CLI, Antigravity, Cursor and the
other hosts on your machine. No sudo, nothing written outside $HOME. Restart
your agent host afterwards and run hephaestus doctor.
curl -fsSL https://raw.githubusercontent.com/agentlas-ai/Agentlas-OS/main/scripts/install-all-runtimes.sh | bash
Agentlas Desktop — the app, if you prefer a window to a terminal.
curl -fsSL https://agentlas.cloud/install.sh | bash # macOS · Linux
irm https://agentlas.cloud/install.ps1 | iex # Windows PowerShell
Downloads the latest signed release, checks its sha512, and installs without sudo or an admin prompt · download manually · Agentlas Desktop is open source — view the repo
Desktop and Agentlas OS together
curl -fsSL https://agentlas.cloud/install.sh | bash -s -- --with-engine
$env:AGENTLAS_WITH_ENGINE="1"; irm https://agentlas.cloud/install.ps1 | iex
What no platform can do
No agent platform routes hundreds or thousands of agents efficiently. Agentlas is the only place that staffs a team of domain specialists onto any job for the fewest tokens.
It works because the agent spec and the protocol for calling an agent are both standardised.
- A2A v1.0, as written. Agent cards are read and written at the standard
/.well-known/agent-card.json. Import, export and caller gating all sit on the standard. An external agent is callable only when its capability is aligned and its identity is verified by a signed card. - One task, 20 candidates: 40,873B → 10,087B. Not fewer candidates — a compressed row. The odds of the right agent being on the list stay put while the tokens drop to a quarter.
- Measured across 849 local profiles and 188 published listings.
This is where most tools stop
| No task split | Task split | |
|---|---|---|
| No specialist sourcing | one chat handles the lump | same model, different prompt |
| Specialist sourcing | pick one from a store | Agentlas |
Some tools do split the work. But what takes each task is the same model with a different prompt. Some places do sell specialists. But there is no split there — you pick one and use it.
Similar plugs
Reverse engineer anything with agents, from app behavior down to native binaries.
Fair-code workflow automation platform with native AI capabilities. Combine visual building with custom code, self-host or cloud, 400+ integrations.
Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype…
User-friendly AI Interface (Supports Ollama, OpenAI API, ...)
A cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Grok Build & Hermes Agent. Only official website: ccswitch.io