Skip to content

PYPI.md

Latest commit

 

History

History
142 lines (103 loc) · 4.12 KB

PYPI.md

File metadata and controls

142 lines (103 loc) · 4.12 KB

OpenAURA logo

PyPI Python CI CodeQL OpenSSF Scorecard Coverage License

OpenAURA is a CI-native Python CLI that generates recurring project briefs from delivery signals and writes markdown reports into aura-docs/.

This page is the technical package reference. The human/product README lives in the GitHub repository.

Requirements

  • Python 3.11 or newer
  • A model provider key for Anthropic or OpenAI
  • At least one signal connector, currently GitHub or Azure DevOps

The project CI verifies Python 3.11, 3.12, 3.13, and 3.14.

Installation

python -m pip install open-aura

The package installs the aura console command.

Configuration

Create aura.config.yml at the root of the repository where briefs should be written:

project: "my-project"
trigger: weekly
schedule: "friday-5pm"
model: "anthropic:claude-sonnet-4-6"

signals:
  github:
    repo: "owner/repository"
    token_env: GITHUB_TOKEN
    default_branch: main

kpis:
  - throughput
  - cycle_time
  - blockers
  - bug_count

custom_kpis: []

output:
  folder: "aura-docs"

Supported trigger values are:

  • weekly
  • on-merge
  • both

The schedule field is documentation for humans. The actual cron schedule belongs in your CI workflow.

Environment Variables

OpenAURA reads secret values from environment variables. The config file stores only the variable names.

Purpose Default variable
Anthropic models ANTHROPIC_API_KEY
OpenAI models OPENAI_API_KEY
GitHub connector GITHUB_TOKEN
Azure DevOps org AZURE_DEVOPS_ORG
Azure DevOps project AZURE_DEVOPS_PROJECT
Azure DevOps token AZURE_DEVOPS_TOKEN

Model strings beginning with anthropic: require ANTHROPIC_API_KEY. Model strings beginning with openai: require OPENAI_API_KEY.

CLI

aura validate
aura run
aura run --dry-run
aura run --trigger on-merge
aura manifesto

aura validate checks the config and required environment variables without making API or LLM calls.

aura run gathers signals, scores the project state, summarizes the brief, and writes markdown to the configured output folder.

aura run --dry-run prints the structured brief JSON instead of writing a file.

GitHub Actions

Install OpenAURA in a workflow with pip:

- uses: actions/setup-python@v6
  with:
    python-version: "3.13"

- run: python -m pip install open-aura
- run: aura validate
- run: aura run
  env:
    GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
    ANTHROPIC_API_KEY: ${{ secrets.ANTHROPIC_API_KEY }}
    OPENAI_API_KEY: ${{ secrets.OPENAI_API_KEY }}

The repository includes starter workflow templates for weekly, on-merge, and combined trigger modes.

Output

By default, briefs are written to:

aura-docs/

Each brief is markdown, intended to be committed back to the same repository by CI.

Links