Skip to main content

MixerSystem workflow engine package

Project description

MixerSystem

MixerSystem is a Python package that runs AI-powered workflows for task creation, planning, implementation, rule updates, and documentation — driven by the context you define in a central .mixer/ folder in your project.

The workflows are built around agentic tools such as Claude Code, Gemini CLI, and Codex CLI.

You configure your project's .mixer/ folder with instructions that shape how each workflow behaves, run mixersystem sync, and MixerSystem generates runtime config and a Claude Code skill wired to your project context. From there, you can use the wired Mixer skill in Claude Code to start running these workflows.

See the details of provided workflows at the bottom of this doc.

Setup

pip install mixer-system
mixersystem sync

To target a different project: mixersystem sync --project_root=/path/to/project

After upgrading (pip install -U mixer-system), re-run mixersystem sync to refresh generated files.

sync generates: .mixer/settings.json, .mixer/README.md, .claude/skills/mixer/SKILL.md

Optional: Linear Integration

Set LINEAR_API_KEY in your project .env and configure .mixer/settings.json:

{ "linear": { "team_prefix": "IOSAI", "team_id": "" } }

Workflows

MixerSystem provides five workflows built around fundamental actions. The core delivery sequence is task → plan → output → update → lesson Each action produces a primary artifact in the task's work folder.

Task

Creates and configures task.md — the starting point for any work item. You provide a task ID, title, and description. MixerSystem routes the task to the right modules based on your project context (or you specify modules explicitly). Supports syncing tasks to and from Linear.

Plan

Produces plan.md from the task. An agentic builder drafts the plan using your project context, then a multi-agent review cycle (architecture, completeness, rules) checks and revises it. Supports parallel branch planning — multiple builders draft independently, then a merger combines the best parts. You can choose which provider runs the builder (Claude, Gemini, or Codex).

Output

Implements the plan. A builder agent writes the code, then a tester agent validates it in a build/test loop (up to a configurable number of iterations). The loop continues until tests pass or the iteration limit is reached. Produces module.md with build status. Supports auto-commit and auto-merge when running inside a git repo.

Update

Creates or updates doc.md — project documentation. Like rules, docs are self-referential: there are docs for how to create docs.

Lesson

Creates or updates rule.md — the rules that govern how builders work. Rules are self-referential: there are rules for how to create rules. The system governs itself using the same mechanisms it exposes to the user.

.mixer/ Configuration

Project-level instructions live in the .mixer/ folder at the project root. Module-specific instructions live under .mixer/modules/, with one folder per module:

.mixer/
  modules/
    cli-a1b2c3/
      task.md       # how to create tasks for this module
      plan.md       # how to create plans for this module
      module.md     # how to build modules for this module
      rule.md       # how to create/update rules for this module
      doc.md        # how to create/update docs for this module
    generator-d4e5f6/
      ...

Each module in the codebase carries a doc.md with a UUID in its frontmatter — the stable identity anchor. The folder name in .mixer/modules/ is <name>-<uuid>. Modules can be renamed or moved without breaking their configuration.

Workflows resolve context by combining project-level instructions with module-specific instructions matched by UUID.

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

mixer_system-0.1.6.tar.gz (34.9 kB view details)

Uploaded Source

Built Distribution

If you're not sure about the file name format, learn more about wheel file names.

mixer_system-0.1.6-py3-none-any.whl (46.5 kB view details)

Uploaded Python 3

File details

Details for the file mixer_system-0.1.6.tar.gz.

File metadata

  • Download URL: mixer_system-0.1.6.tar.gz
  • Upload date:
  • Size: 34.9 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.2

File hashes

Hashes for mixer_system-0.1.6.tar.gz
Algorithm Hash digest
SHA256 488ceba21f2b67d1e5741a8c3ff1e0b811891edc0fb07f57abba19477299d9d7
MD5 04c54a1b35041cab0a23cb57731f54c7
BLAKE2b-256 4abba3cafe7d9d612c499045ad1a28443eed85d46c97357654c522d1496a53db

See more details on using hashes here.

File details

Details for the file mixer_system-0.1.6-py3-none-any.whl.

File metadata

  • Download URL: mixer_system-0.1.6-py3-none-any.whl
  • Upload date:
  • Size: 46.5 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: twine/6.2.0 CPython/3.12.2

File hashes

Hashes for mixer_system-0.1.6-py3-none-any.whl
Algorithm Hash digest
SHA256 7da1340dfbcc41a2f08e651602d49bd242632554d60ed01a810732bfc12bbd29
MD5 bc091c740722293b6942890e8bc0d356
BLAKE2b-256 4d10503793a1771e89d4822aa656b80ec11b7ede1a3f0915d2baa0e44df331de

See more details on using hashes here.

Supported by

AWS Cloud computing and Security Sponsor Datadog Monitoring Depot Continuous Integration Fastly CDN Google Download Analytics Pingdom Monitoring Sentry Error logging StatusPage Status page