Helix is a powerful self-learning AI agent with memory, skills, and the ability to recall tools. It learns from successful tasks and creates reusable skills for the future.
This repository also configures its own agents. See what Holix tells them →
Skill Claude CodeCodex
A release workflow for the Holix GitHub repository. It prepares a pull request, checks that automated tests pass, merges it, and creates the next numbered release.
Skill Claude Code
Schedule recurring agent tasks via Holix built-in gateway cron (not crontab or custom scripts).
Skill Claude Code
Author and manage Holix agent drop-in extensions so the agent can extend itself without editing Holix core. Use when the user asks to write an extension, add tools, register slash commands, list extensions, disable a broken extension, or self-modify agent capabilities. Invoke via /holix-extensions.
Skill Claude Code
Implement an SDD change from tasks.md — always ask self vs subagents first, then execute only after apply mode is set.
Skill Claude Code
Archive a completed SDD change — merge delta specs into main openspec/specs and move change to archive.
Skill Claude Code
Create Spec-Driven Development changes (OpenSpec-style) — multi-project openspec, understanding gate, assigned tasks before coding.
Skill Claude Code needs its repo
Run frontend (and FE+BE) apps in Holix Studio: always bind 0.0.0.0, always open Studio preview links in chat, always prefer docker-compose with an nginx reverse proxy that joins frontend and backend. Use when starting Vite/Next/Nuxt/React, SPA+API monorepos, or any web app the user should open in Studio Browser.…
Skill Claude Code
Full Holix sub-agent lifecycle: spawn, monitor, collect results, answer askuser questions, approve tool confirmations, and delegate to external CLIs (holix launch). Use when the user asks about subagents, /subagent-spawn, delegatetosubagent, background workers, sub-agent questions, /subagent-reply, holix launch +…
Skill Claude Code
A guide for using a separate web-researcher agent. It applies only when someone explicitly asks that agent to research something.
Skill Claude Code
A self-diagnosis skill for an AI assistant that runs when the user says the assistant is wrong or asks it to check itself. It reads a diagnostic report before explaining what happened.
Skill Claude CodeCodex
Generate images and short videos from a prompt, or from user-uploaded reference photos (edit, combine, animate).
Skill Claude Code
Optional. Cosine similarity of texts you already have, via textsimilarity. Not a classifier and not Holix memory search.
Skill Claude Code
Optional. Closed-set choice, an ordinal score, or a yes/no probability via systemonedecide. The decision model does not write the user-facing answer.
Skill Claude Code
Run a command on another machine with sshexec. Use when the user names a server, host, IP, or asks to SSH. Do not assemble ssh, scp, or sftp in the local shell.
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: