agentic-stack

agentic-stack

A portable .agent/ folder containing memory, skills, and protocols that preserves knowledge when switching between multiple coding agent tools.

Agent SkillDevOpen source
Type
Agent Skill
Open source
Yes
GitHub Stars
★ 2.1k
Source
skill-github

Overview

agentic-stack is a portable .agent/ folder containing memory, skills, and protocols. It can be inserted into various coding agent tools such as Claude Code, Cursor, Windsurf, OpenCode, OpenClaw, Hermes, or DIY Python, preserving its knowledge when switching tools. It also includes a local data layer that monitors activity across all agent tools, cron runs, active agents, token/cost estimates, and KPI summaries. Additionally, it can convert approved, edited runs into local flywheel artifacts such as audit logs, context cards, evaluation cases, training-ready JSONL, and readiness metrics.

Capabilities

  • Maintains memory and skills across tools
  • Local data layer monitors agent activity
  • Generates local dashboard exports
  • Exports approved run records

Use cases

Switching between multiple coding agent toolsMonitoring and managing multiple agent toolsGenerating and exporting local data reportsCreating and saving visual canvases

Setup

Requires: Python 环境
Install via `pip install agentic-stack`, then initialize your project with `agentic-stack init`.

This information was compiled by AI from public sources and may contain inaccuracies — please refer to the source.

FAQ

How do I switch between tools?

Simply insert the .agent/ folder into the new tool.

How do I enable the tldraw skill?

Set `tldraw.enabled: true` in the config file and merge the MCP configuration.

Related skills