Attune AI Action

AI-powered code review and release prep for your CI/CD pipeline, powered by Attune AI and Claude.

Smart tier routing automatically optimizes LLM costs (34-86% savings) while maintaining quality.

Quick Start

Code Review on PRs

name: Attune Code Review
on: [pull_request]

jobs:
  review:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 2

      - uses: Smart-AI-Memory/attune-ai-action@v1
        with:
          workflow: code-review
          anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}

Release Prep on Tags

name: Attune Release Prep
on:
  push:
    tags: ['v*']

jobs:
  release-prep:
    runs-on: ubuntu-latest
    steps:
      - uses: actions/checkout@v4
        with:
          fetch-depth: 0

      - uses: Smart-AI-Memory/attune-ai-action@v1
        with:
          workflow: release-prep
          anthropic_api_key: ${{ secrets.ANTHROPIC_API_KEY }}
          fail_on_critical: 'true'

Inputs

Input	Required	Default	Description
workflow	Yes	-	code-review or release-prep
anthropic_api_key	Yes	-	Your Anthropic API key
python_version	No	3.12	Python version
attune_version	No	latest	Attune AI version
tier_strategy	No	auto	Cost tier: auto, cheap, capable, premium
fail_on_critical	No	true	Fail action on critical issues
config	No	-	Path to attune config file

Outputs

Output	Description
report	Path to generated report
summary	Brief summary of findings
issues_found	Number of issues found
cost_saved	Estimated cost savings from tier routing

Cost Optimization

Attune AI routes prompts through smart tiers (CHEAP -> CAPABLE -> PREMIUM), only escalating when complexity requires it. In CI, this means routine reviews use cheaper models while critical security findings get premium attention — saving 34-86% vs. using the most expensive model for everything.

Set tier_strategy to control this:

• auto (default) — Attune decides based on complexity
• cheap — minimize cost, good for quick checks
• capable — balanced quality and cost
• premium — maximum quality, use for release-critical reviews

License

Apache 2.0 — same as Attune AI.