Autonomous AI orchestration with persistent memory, self-coordinating agent orgs, and a codebase knowledge graph and more.
This repository also configures its own agents. See what monomind tells them →
Skill Claude CodeCodex needs its repo
UI QA and browser automation with the native monomind browse CDP client: discovery, user flows, performance profiling, and pass/fail/warn reports.
Skill Claude CodeCodex
Self-learning version control for AI agents with ReasoningBank intelligence and multi-agent coordination. Wraps the separate agentic-jujutsu npm package.
Skill Claude CodeCodex
GitHub workflows for monomind projects: issues, PRs, releases, repo structure, multi-package sync, Actions. Uses the gh CLI; monomind has no github command.
Skill Claude Code needs its repo
Set up and use monomind hooks: pre/post edit, command and task hooks, sessions, routing, pattern logging, background workers, settings.json and Git hooks.
Skill Claude CodeCodex
Mastermind access — bookkeeping only — not enforced by the Org Runtime. Records member roles, permission grants, invite tokens and join requests in a side file nothing checks.
Skill Claude CodeCodex
Mastermind activity — an org's activity feed (up to 200 recent issue, project, goal, agent and routine events) with event-type filters and actor/entity names.
Skill Claude CodeCodex
Mastermind adapter-manager — global adapter catalog: list built-in and npm adapters, enable/disable them, reinstall from npm, add or remove custom HTTP adapters.
Skill Claude CodeCodex
Mastermind adapters — install, enable, disable, reload, and remove LLM adapter plugins per org. Supports claude-local, gemini-local, codex-local, cursor, hermes, http, and custom adapters.
Skill Claude CodeCodex
Mastermind agent-detail — deep per-agent inspection: show config, run history, budget usage, heartbeat status, assigned skills, and reset/reconfigure a single agent within an org.
Skill Claude CodeCodex needs its repo
Shared utility: ranks agents from .monomind/registry.json for a task and category, so mastermind skills pick agents instead of hardcoding them.
Skill Claude CodeCodex
Mastermind agents — list, inspect, hire, and remove agents in an org, and pause or resume the org. Shows each role, its adapter config, and the org's live runtime status.
Skill Claude CodeCodex
Mastermind approval-detail — inspect and resolve one approval request from an org's agents via monomind org approvals, org approve and org deny.
Skill Claude CodeCodex
Mastermind backup — create, list and restore timestamped tarball backups of all of an org's data files.
Skill Claude CodeCodex
Mastermind bootstrap — one-time org initialization that primes the CEO/boss agent with org context, goal hierarchy, and a signed invite token. Run after createorg before first runorg.
Skill Claude CodeCodex
Mastermind budgets — view, set and track the spend caps the Org Runtime enforces per role, compare them with recorded spend, and hot-reload a running org.
Skill Claude CodeCodex
Mastermind createorg — design an autonomous agent org and save it as .monomind/orgs/ .json for monomind org run/serve; --schedule for daemon-scheduled orgs.
Skill Claude CodeCodex
Root-cause debugging protocol to run before any fix: test failures, bugs, unexpected behavior, build failures, performance regressions.
Skill Claude CodeCodex
Delegation capability injected into spawned agent prompts so subagents can delegate further. Never invoked directly.
Skill Claude CodeCodex
MUST use before creative work (features, components, behavior changes): explores intent, requirements and design with the user before implementation.
Skill Claude CodeCodex
Mastermind diagnose — find why agent work stalled, looped or went too deep: locate the stop-point in the issue tree and get a fix plan approved before code changes.
Skill Claude CodeCodex
Mastermind env — audit an org's runtime environment (LLM provider, memory backend, agent JWT, logging, storage) and flag missing or misconfigured values.
Skill Claude CodeCodex
Mastermind environments — bookkeeping only — not enforced by the Org Runtime. Records execution environments (local, SSH, sandbox) in a side file; agents run where the runtime runs them.
Skill Claude CodeCodex
Load a written implementation plan, review it critically, execute all tasks step by step, and hand off to mastermind:review when complete.
Skill Claude CodeCodex
Mastermind export — export an org to a tar.gz or zip archive with file selection, or import one with a collision strategy and adapter override.
At most 3 mods per repository are shown here, and a mod shipped inside a plugin is left to that plugin's page — the rest are on their repository pages: