Configuration Reference

Complete reference for all configuration options available in the AI Workflow Failure Summary Action.

Input Parameters

Required Inputs

github-token

• Type: string
• Required: Yes
• Default: ${{ github.token }}
• Description: GitHub token for accessing workflow logs and creating issues

github-token: ${{ secrets.GITHUB_TOKEN }}

LLM Provider Configuration

llm-provider

• Type: string
• Required: No
• Default: openai
• Options: openai, azure-openai, github-models, anthropic
• Description: Which LLM provider to use for analysis

llm-provider: 'openai'

OpenAI Configuration

openai-api-key

• Type: string
• Required: When using OpenAI provider
• Description: OpenAI API key

openai-api-key: ${{ secrets.OPENAI_API_KEY }}

openai-model

• Type: string
• Required: No
• Default: gpt-4o-mini
• Options: Any OpenAI chat model (e.g., gpt-4o, gpt-4o-mini, gpt-4-turbo)
• Description: OpenAI model to use for analysis

openai-model: 'gpt-4o'

Azure OpenAI Configuration

azure-openai-api-key

• Type: string
• Required: When using Azure OpenAI provider
• Description: Azure OpenAI API key

azure-openai-api-key: ${{ secrets.AZURE_OPENAI_API_KEY }}

azure-openai-endpoint

• Type: string
• Required: When using Azure OpenAI provider
• Description: Azure OpenAI endpoint URL

azure-openai-endpoint: 'https://your-resource.openai.azure.com'

azure-openai-deployment

• Type: string
• Required: When using Azure OpenAI provider
• Description: Azure OpenAI deployment name

azure-openai-deployment: 'gpt-4o-deployment'

GitHub Models Configuration

github-models-token

• Type: string
• Required: When using GitHub Models provider
• Description: GitHub PAT with model access (can use GITHUB_TOKEN)

github-models-token: ${{ secrets.GITHUB_TOKEN }}

github-models-model

• Type: string
• Required: No
• Default: gpt-4o
• Options: gpt-4o, gpt-4o-mini, claude-3-5-sonnet, etc.
• Description: GitHub Models model to use

github-models-model: 'gpt-4o'

Anthropic Configuration

anthropic-api-key

• Type: string
• Required: When using Anthropic provider
• Description: Anthropic API key

anthropic-api-key: ${{ secrets.ANTHROPIC_API_KEY }}

anthropic-model

• Type: string
• Required: No
• Default: claude-3-5-sonnet-20241022
• Options: claude-3-5-sonnet-20241022, claude-3-opus-20240229, etc.
• Description: Anthropic model to use

anthropic-model: 'claude-3-5-sonnet-20241022'

Analysis Configuration

max-log-lines

• Type: string
• Required: No
• Default: 500
• Description: Maximum number of log lines to analyze per failed job

max-log-lines: '1000'

Token Limits

Higher values provide more context but use more tokens. Consider your LLM's context window and costs.

Issue Creation

create-issue

• Type: string (boolean)
• Required: No
• Default: false
• Options: true, false
• Description: Whether to create a GitHub issue with the analysis

create-issue: 'true'

issue-label

• Type: string
• Required: No
• Default: ai-analysis
• Description: Label to add to created issues

issue-label: 'automated-analysis'

Custom Prompts

custom-system-prompt

• Type: string
• Required: No
• Description: Custom system prompt or path to prompt file

custom-system-prompt: '.github/prompts/system-prompt.md'
# or inline:
custom-system-prompt: 'You are a Python expert analyzing pytest failures...'

custom-user-prompt

• Type: string
• Required: No
• Description: Custom user prompt template or path to prompt file

custom-user-prompt: '.github/prompts/user-prompt.md'

See Custom Prompts for details.

Output Parameters

summary

• Type: string
• Description: AI-generated summary of workflow failures

- id: analyze
  uses: ianlintner/ai_summary_action@v1
  # ... inputs

- name: Use Summary
  run: echo "${{ steps.analyze.outputs.summary }}"

failed-jobs

• Type: string (JSON array)
• Description: Array of failed job names

- name: List Failed Jobs
  run: echo "${{ steps.analyze.outputs.failed-jobs }}"

issue-url

• Type: string
• Description: URL of created issue (only when create-issue is true)

- name: Link to Issue
  if: steps.analyze.outputs.issue-url
  run: echo "Issue created: ${{ steps.analyze.outputs.issue-url }}"

Complete Configuration Examples

Minimal Configuration

- uses: ianlintner/ai_summary_action@v1
  with:
    github-token: ${{ secrets.GITHUB_TOKEN }}
    openai-api-key: ${{ secrets.OPENAI_API_KEY }}

Recommended Configuration

- uses: ianlintner/ai_summary_action@v1
  with:
    github-token: ${{ secrets.GITHUB_TOKEN }}
    llm-provider: 'openai'
    openai-api-key: ${{ secrets.OPENAI_API_KEY }}
    openai-model: 'gpt-4o-mini'
    max-log-lines: '500'
    create-issue: 'true'
    issue-label: 'ci-failure'

Advanced Configuration

- uses: ianlintner/ai_summary_action@v1
  with:
    github-token: ${{ secrets.GITHUB_TOKEN }}
    llm-provider: 'anthropic'
    anthropic-api-key: ${{ secrets.ANTHROPIC_API_KEY }}
    anthropic-model: 'claude-3-5-sonnet-20241022'
    max-log-lines: '1000'
    create-issue: 'true'
    issue-label: 'automated-analysis'
    custom-system-prompt: '.github/prompts/system-prompt.md'
    custom-user-prompt: '.github/prompts/user-prompt.md'

Environment Variables

The action respects these environment variables:

• GITHUB_WORKSPACE: Used to resolve relative paths for custom prompts
• GITHUB_TOKEN: Can be used implicitly via ${{ github.token }}

Permissions

Required workflow permissions:

permissions:
  actions: read       # Required: Read workflow logs
  contents: read      # Required: Access repository
  issues: write       # Optional: Only if create-issue is true

Best Practices

Cost Optimization

# Use smaller, cheaper models for most analysis
openai-model: 'gpt-4o-mini'
max-log-lines: '500'

# Or use free GitHub Models
llm-provider: 'github-models'
github-models-token: ${{ secrets.GITHUB_TOKEN }}

Quality Optimization

# Use more powerful models for better analysis
llm-provider: 'anthropic'
anthropic-model: 'claude-3-5-sonnet-20241022'
max-log-lines: '1000'
custom-system-prompt: '.github/prompts/detailed-analysis.md'

Security

# Never hardcode API keys
openai-api-key: ${{ secrets.OPENAI_API_KEY }}  # ✅ Good

# Don't do this:
# openai-api-key: 'sk-...'  # ❌ Bad

Troubleshooting Configuration