Show HN: Topos – Structural code quality metrics for agent-written programs

Category: devtools

Tags: code-quality, code-analysis, ai-agents, static-analysis, mcp-server

Score: 7.0/10 (Innovation: 6, Technical: 8, Documentation: 8, Utility: 6)

Topos provides structural code quality metrics (simplicity, composability, security) for evaluating code generated by AI agents, going beyond unit tests. It combines AST/CFG analysis, module dependency graphs, and code property graphs to award medals, and integrates with MCP servers for iterative agent feedback loops. Its focus on agent-written code and the three-pillar medal system is a novel approach to an emerging problem.

Target audience: AI/ML engineers, backend devs, devops

Repository: https://krv.ai/field-notes/evaluating-code-generation · Python · BSD-3-Clause · 6 stars

View on Hacker News