SentinelCI

SentinelCI solves the problem of slow CI/CD pipelines that run all tests for every change. We learned: Building AI agents with semantic code analysis, implementing intelligent test selection algorithms, and creating AI-driven risk scoring for CI/CD workflows. How we built it: Using Python, OpenRouter API (Claude/GPT-4), GitLab API, Tree-sitter for AST parsing, and semantic embeddings for intelligent test selection. The system has 9 phases from repository analysis to automated CI/CD gating. Challenges: Balancing accuracy vs performance in semantic analysis, handling complex codebases with dynamic dependencies, and creating meaningful risk scores that align with real-world scenarios.

Built With

• built-with:-python
• chromadb
• claude-ai
• gitlab
• gitlab-api
• gpt-4
• networkx
• openrouter-api
• pytest
• sentence-transformers
• tree-sitter