GitHub - yksanjo/agentic-ci: CI that understands code, not just executes it. LLM-powered CI pipeline that predicts tests, explains failures, and cuts debug time by 40%.

❌ Traditional CI Failure:
┌─────────────────────────────────────────────────────────────┐
│  ❌ Test Suite Failed                                       │
│  ─────────────────────────────────────────────────────────  │
│  tests/api/test_users.py::test_create_user ............ FAIL│
│  tests/api/test_orders.py::test_place_order .......... FAIL│
│  tests/integration/test_payment.py::test_refund ...... FAIL│
│  tests/unit/test_utils.py::test_format_date ......... FAIL│
│                                                             │
│  47 tests passed, 4 failed                                  │
│  Log output: 2,847 lines                                    │
│                                                             │
│  [Scroll through logs manually to find the issue...]        │
└─────────────────────────────────────────────────────────────┘
Time to resolution: 45 minutes

✅ Agentic CI Failure Analysis:
┌─────────────────────────────────────────────────────────────┐
│  🔍 Smart Failure Detected                                  │
│  ─────────────────────────────────────────────────────────  │
│                                                             │
│  📍 Root Cause: src/api/users.py:142                        │
│     └─ Database transaction rollback missing in error path  │
│                                                             │
│  🔗 Affected Components:                                    │
│     • User creation API                                     │
│     • Order placement (cascading failure)                   │
│                                                             │
│  💡 Suggested Fix:                                          │
│     Add session.rollback() in exception handler at          │
│     src/api/users.py:147                                    │
│                                                             │
│  📚 Similar Past Failures:                                  │
│     • PR #2842 (2 weeks ago) - same pattern                 │
│                                                             │
│  ⚡ Only 12 tests need re-run (not 51)                      │
└─────────────────────────────────────────────────────────────┘
Time to resolution: 8 minutes

# Clone the repository
git clone https://github.com/yksanjo/agentic-ci.git
cd agentic-ci

# Create virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

# Pull a code-focused LLM (if using Ollama)
ollama pull codellama:7b

# Start the server
uvicorn api.main:app --reload --port 8080

# Or use the module directly
python -m api.main

# Get a diff
git diff HEAD~1 > changes.diff

# Analyze it
curl -X POST http://localhost:8080/analyze \
  -H "Content-Type: application/json" \
  -d '{
    "diff": "'"$(cat changes.diff)"'",
    "branch": "feature/new-feature",
    "author": "developer@example.com"
  }'

curl -X POST http://localhost:8080/predict/tests \
  -H "Content-Type: application/json" \
  -d '{
    "changed_files": [
      {"path": "src/api/users.py", "additions": 50, "deletions": 10}
    ]
  }'

curl -X POST http://localhost:8080/explain/failure \
  -H "Content-Type: application/json" \
  -d '{
    "failure_log": "AssertionError: Expected 5 but got 4\n  at test_calculator.py:42"
  }'

curl http://localhost:8080/optimizer/report?days=7

# LLM settings
llm:
  provider: "ollama"  # or openai, anthropic
  model: "codellama:7b"
  base_url: "http://localhost:11434"

# Risk assessment weights
risk:
  weights:
    file_criticality: 0.30
    change_complexity: 0.25
    historical_risk: 0.20
    semantic_risk: 0.15
    dependency_risk: 0.10

# Flaky test thresholds
optimizer:
  flaky_threshold: 0.10
  quarantine_threshold: 0.30

# LLM Configuration
export LLM_PROVIDER=ollama
export LLM_MODEL=codellama:7b
export LLM_BASE_URL=http://localhost:11434
export AGENTIC_CI_API_KEY=your-api-key  # For OpenAI/Anthropic

# API Configuration
export API_HOST=0.0.0.0
export API_PORT=8080

# Storage
export PATTERN_STORE_PATH=./data/patterns

agentic-ci/
├── agentic_ci/
│   ├── __init__.py
│   └── core/
│       ├── analyzer.py      # Change analysis orchestrator
│       ├── predictor.py     # Test prediction engine
│       ├── explainer.py     # Failure root cause analysis
│       ├── optimizer.py     # CI optimization & flaky test management
│       ├── llm_client.py    # Multi-provider LLM client
│       ├── risk_scorer.py   # Risk assessment
│       └── pattern_store.py # Historical pattern storage
├── api/
│   └── main.py              # FastAPI REST API
├── config/
│   └── config.yaml          # Configuration
├── data/                    # Persisted patterns
└── tests/                   # Test suite

name: Agentic CI

on: [push, pull_request]

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

      - name: Get diff
        run: git diff ${{ github.event.before }}..${{ github.sha }} > changes.diff

      - name: Analyze changes
        run: |
          curl -X POST $AGENTIC_CI_URL/analyze \
            -H "Content-Type: application/json" \
            -d '{"diff": "'"$(cat changes.diff)"'"}'

agentic-analyze:
  stage: test
  script:
    - git diff HEAD~1 > changes.diff
    - |
      curl -X POST $AGENTIC_CI_URL/analyze \
        -H "Content-Type: application/json" \
        -d '{"diff": "'"$(cat changes.diff)"'"}'

# Install dev dependencies
pip install -r requirements.txt

# Run tests
pytest tests/ -v

# Format code
black agentic_ci/
isort agentic_ci/

# Type checking
mypy agentic_ci/

# Lint
ruff check agentic_ci/