The agent harness that remembers your work — so you get the assistant you actually wanted.
Works with Claude Code + Antigravity — see compatibility
AgentM is built to learn how you work — and, in time, to know it better than you do. It helps you remember: the decisions, the open threads, the conventions, the step you always forget before a release. It knows how the work should be done — it is opinionated, with a sound answer ready even for the things you didn't think to ask. And when it hits something it doesn't know, it works it out and keeps what it learns for the next time.
Said differently, AgentM combines a custom harness with persistent memory and self-improvement — it dreams and learns between sessions. Its customization system extends those abilities through plugins that enable long-running, nuanced development workflows, automated project management, and more. Its opinion system is designed to layer in personas focused on a specific kind of work — activation is the next build slice.
AgentM works best paired with crickets — the toolkit of plugins (capabilities, skills, hooks, and sub-agents) that make it even more useful.
AgentM installs alongside your coding agent — Claude Code or Antigravity — so you can be up and running in a few minutes. See requirements.
Install AgentM and crickets with the recommended configuration and a Google Drive vault:
# Point the vault at a Google Drive folder named "Agent" (the recommended default)
export MEMORY_VAULT_PATH="<your-google-drive>/Agent"
# Install AgentM for every project on this machine (user scope) + verification hooks
bash ~/agentm/install.sh --hooks --scope user
# Add the crickets plugins (Claude Code + Antigravity)
curl -fsSL https://raw.githubusercontent.com/alexherrero/crickets/main/bootstrap.sh | bashMore on the available configurations here.
The wiki covers everything there is to know about AgentM. A few links to get you started.
- Why we built it — the problem, the solution and the reason.
- Architecture — how it's made.
- Reference — fields, flags, schemas and more.
Note
Latest release: v5.13.0. The persona roster grows from 2 to 11 authored manifests, with a re-audited effort-routing binding and content-refresh's first named consumer — Architecture-Governance Wave D.
Contributions are welcome — see CONTRIBUTING.md for information on how to contribute to AgentM.
Multi-licensed so each layer carries the license that fits it: code under Apache-2.0, content — docs, prompts, prose, and other .md — under CC-BY-4.0, and the "AgentM" name and brand under trademark. Prompt text embedded in a code file counts as content, even though it lives there. Both licenses allow commercial use and derivative works with attribution — see NOTICE for the attribution string.