LLM Runtime Architecture
SkillAI & modelsThis skill helps your AI design how a single agent core can run across Codex, Claude Code, Gemini CLI, Cursor, and other AGENTS.md-compatible tools. Once added, your AI can work out a runtime design where the same core behaves consistently no matter which coding tool runs it. It comes from the agentlas-ai/agentlas-os repository.
Available today. Use it from your connected AI after setup.
No other account needed.
After adding it, ask your AI to draft a runtime design for your agent core that covers Codex, Claude Code, Gemini CLI, and Cursor.
Then ask your AI: use the LLM Runtime Architecture skill
What your AI can do with it
- Design one agent core that runs across Codex, Claude Code, Gemini CLI, and Cursor
- Plan how the same core stays consistent across AGENTS.md-compatible tools
- Work through runtime design decisions for each supported coding tool
- Outline how agent behavior carries over when moving between these tools
What this skill tells your AI
The instructions your AI receives, as published by agentlas-ai/agentlas-os in skills/llm-runtime-architecture/SKILL.md and read by ahel’s review.
Procedure
- Keep
AGENTS.mdas the canonical behavior contract. - For each runtime, name entry point, global command, adapter files, available tools, memory access, limitations, and verification command.
- Keep adapters thin and point them back to the canonical core.
- Write or repair
.agentlas/global-commands.jsonwhen creating or packaging an agent. - State unsupported capabilities explicitly.
Output
Return a runtime matrix with runtime, entry_point, global_command,
adapter_files, memory_access, limitations, and verification.
Signals
- GitHub stars
- 1k
- Forks
- 103
- Last commit
- Sep 2026
Advanced
- Catalog kind
- skill
- Gateway key
llm-runtime-architecture- Source
- github.com/agentlas-ai/agentlas-os