Fail-closed, hash-audited agent workspaces built from read manifests.
-
Updated
Aug 10, 2026 - Python
Fail-closed, hash-audited agent workspaces built from read manifests.
One workspace for all your AI work — build software, create videos, research ideas, write documents. Self-hosted, persistent across sessions, CLI-agnostic (Claude Code / Codex / Cursor). Plain Markdown, zero install.
OpenClaw workspace and configuration for a personal AI agent.
To associate your repository with the agent-workspaces topic, visit your repo's landing page and select "manage topics."