Skip to main content

Deterministic agent-management framework that implements FRAME governance for AI-assisted software development.

Project description

Haxaml

PyPI version

Haxaml is a deterministic project-governance framework for AI coding workflows.

MCP first. CLI second.

It gives agents one shared source of truth called FRAME so context does not drift across tools.

What it does

  • keeps project truth in structured FRAME files
  • validates that structure before work starts
  • tracks runs, decisions, risks, and completed work
  • exports clean agent-ready prompt files from FRAME
  • exposes MCP tools so IDE agents can call governance directly

FRAME at a glance

File Purpose
.haxaml/facts.yaml project truth
.haxaml/rules.yaml how agents should work
.haxaml/acts.yaml task diary and decisions
.haxaml/expect.yaml expected next runs
.haxaml/map.yaml optional module ownership + impact map

Install

pip install haxaml

or with uv tools:

uv tool install haxaml
uv tool install haxaml-mcp

Quick start

haxaml init
haxaml validate
haxaml context --tokens
haxaml export

Default export writes HAXAML.md.

Optional agent-specific exports:

haxaml export --agent claude
haxaml export --agent codex
haxaml export --agent copilot
haxaml export --agent cursor
haxaml export --agent windsurf
haxaml export --agent gemini

Safe preview before writing:

haxaml export --dry-run --diff-preview --target CLAUDE.md

MCP setup (recommended)

Use uvx in your MCP client config:

{
  "mcpServers": {
    "haxaml": {
      "command": "uvx",
      "args": ["haxaml-mcp"],
      "env": {
        "HAXAML_PROJECT_DIR": "/path/to/project"
      }
    }
  }
}

You can also generate project-local bootstrap config with:

haxaml mcp-bootstrap --mode write --editor generic

Upgrade

haxaml upgrade

Pin a specific version:

haxaml upgrade --to 0.2.0

Why FRAME even when markdown can be smaller?

In some fixtures, raw markdown is smaller than YAML. That is normal.

Haxaml is not only about raw token count. The value is deterministic governance:

  • validation gates
  • compact compiled exports
  • explicit decisions and risks
  • stable multi-agent handoff

Docs

  • ADVANCED.md for full command reference, MCP tool contract, export behavior, and roadmap notes.

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

haxaml-0.2.0.tar.gz (71.5 kB view details)

Uploaded Source

Built Distribution

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

haxaml-0.2.0-py3-none-any.whl (64.3 kB view details)

Uploaded Python 3

File details

Details for the file haxaml-0.2.0.tar.gz.

File metadata

  • Download URL: haxaml-0.2.0.tar.gz
  • Upload date:
  • Size: 71.5 kB
  • Tags: Source
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.8.4

File hashes

Hashes for haxaml-0.2.0.tar.gz
Algorithm Hash digest
SHA256 a9f87ed7f6ad3acd2c0beaf5ef33618ef5ee77b9b86aa21f52484f707a11816b
MD5 ae4320e607c7c96148579f9f964cc9b2
BLAKE2b-256 0984d8358f0becb13a6bfb8ea4b2aec333351807fb558dbae7741931b7549332

See more details on using hashes here.

File details

Details for the file haxaml-0.2.0-py3-none-any.whl.

File metadata

  • Download URL: haxaml-0.2.0-py3-none-any.whl
  • Upload date:
  • Size: 64.3 kB
  • Tags: Python 3
  • Uploaded using Trusted Publishing? No
  • Uploaded via: uv/0.8.4

File hashes

Hashes for haxaml-0.2.0-py3-none-any.whl
Algorithm Hash digest
SHA256 01d9a4778dd3622afe86b7895ee9768d251a65774241bb5dbebab7132d00927a
MD5 87d9fb8ff7617ff0907652cbbf5540bc
BLAKE2b-256 178e368c8cbdc7d5cf63afa74e3fff1a40a9c4810875009e2bfb5cbed110e46f

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