Keep a portable agent workspace layer with agentic-stack
Install and maintain a shared .agent layer so coding agents can carry memory, skills, protocols, dashboards, and bounded loop contracts across multiple agent harnesses.
npx skills add agentskillexchange/skills --skill keep-a-portable-agent-workspace-layer-with-agentic-stack
Use agentic-stack when an operator wants the same project-local agent memory, skills, protocols, and maintenance routines available across Claude Code, Cursor, Windsurf, OpenCode, OpenClaw, Codex, Gemini CLI, or a standalone Python loop. The workflow is to install the .agent layer into a project, attach the relevant harness adapter, run doctor or dashboard checks, upgrade the skeleton safely, and use loop commands only for bounded, audited coding-agent runs. This is not a general agent framework listing; the boundary is portable workspace operations for existing coding-agent harnesses: install, add adapters, audit, upgrade, sync manifests, and run bounded local loops without clobbering project memory.
What this skill actually does
Inputs and prerequisites: Homebrew or git, shell, supported coding-agent harness.
Setup notes: Install with brew tap codejunkie99/agentic-stack https://github.com/codejunkie99/agentic-stack and brew install agentic-stack , then run agentic-stack claude-code , agentic-stack cursor , agentic-stack openclaw , agentic-stack codex , or another supported adapter inside the target project. For clone-based setup, run git clone https://github.com/codejunkie99/agentic-stack.git , cd agentic-stack , and ./install.sh .
Source and verification boundary: use https://github.com/codejunkie99/agentic-stack as the canonical reference before running the workflow; keep commands, API calls, CLI usage, and generated outputs reviewable against that upstream source.
Framework fit: publish this as a Multi-Framework workflow only when the operator can invoke the documented toolchain directly, rather than treating the upstream project as a generic product listing.