agentic-stack
codejunkie99/agentic-stack
Portable .agent/ folder of memory, skills and protocols that moves between coding-agent harnesses
agentic-stack is a portable .agent/ folder holding memory, skills and protocols that plugs into many coding agents, Hermes among them. It keeps what an agent has learned when you switch tools.
What agentic-stack does
agentic-stack keeps one memory-and-skills layer in a project-level .agent/ folder and wires it into whichever coding agent you use. The adapter list includes Claude Code, Cursor, Windsurf, OpenCode, OpenClaw, GitHub Copilot CLI, Gemini CLI, Hermes, Pi Coding Agent, Codex, Autohand Code CLI and Antigravity, plus a plain Python loop. Recall skips lessons that a newer lesson has superseded.
A local data layer lets you watch harness activity, cron runs, active agents, token and cost estimates and daily dashboards. Approved, redacted runs can be turned into local artifacts such as trace records, eval cases and training-ready JSONL without sending telemetry. Version 0.19 adds bounded loops under .agent/loops, where a maker, a deterministic verifier and an independent checker run with finite attempt, runtime and token budgets. The README states the loop supervisor is not an operating-system sandbox.
Key features
- Adapters for Claude Code, Cursor, Windsurf, OpenCode, OpenClaw, Hermes, Codex and others
- Verb-style commands: add, status, doctor, upgrade, remove and dashboard
- TUI dashboard plus a beta local web dashboard called mission-control
- Local data layer for harness activity, cron runs and token and cost estimates
- Bounded loops with owned Git worktrees, deny-path gates and resumable checkpoints
- Memory export and import through a transfer wizard
When to use it
- Switch between Hermes and another coding agent without losing accumulated lessons
- Run several adapters in one repository, for example Claude Code and Cursor together
- Watch cost and activity across a mixed set of agents from one dashboard
- Run a bounded CI-fixing loop against a failing test
Who it is for: Developers who use more than one coding agent and want shared memory and skills that survive switching tools.
How it fits with Hermes Agent
Hermes is one of the listed adapters, so agentic-stack supports Hermes Agent among many other agents rather than being built only for it.
How to install agentic-stack
These commands are copied from the project's README. Check the repository for the latest steps before you run them.
brew tap codejunkie99/agentic-stack https://github.com/codejunkie99/agentic-stack
brew install agentic-stack FAQ
What is agentic-stack?
agentic-stack is a portable .agent/ folder containing memory, skills and protocols for coding agents. It installs per project and carries its knowledge across harnesses.
Does agentic-stack work with Hermes Agent?
Yes. Hermes is one of the supported adapters. After installing, run agentic-stack hermes from your project directory.
Is agentic-stack free and open source?
Yes, it is open source under Apache-2.0, and it keeps its data layer local without sending telemetry.
Similar memory for Hermes Agent
All memoryLocal-first Markdown memory kit for agents, with a Hermes Agent memory provider
mnemosyne-oss MnemosyneHermes-first AI memory layer in one SQLite file with one pure-Python dependency and no cloud
moorcheh-ai MemantoMemory agent that curates, reconciles and shares long-term memory across multiple AI agents
MemTensor MemmyLocal memory hub and personal agent that shares one memory across Claude Code, Codex, OpenClaw and Hermes
activeloopai HivemindShared agent memory that turns session traces into reusable skills, with a Hermes install option
ClaudioDrews Memory OSSeven-layer local memory system for Hermes Agent with Qdrant, structured facts and a curated wiki
Related guides: SOUL.md for Hermes Agent: what it is and how to write one · Run multiple Hermes agents with profiles