Why it was accepted
The page is clearly about an AI-agent evaluation project with concrete evidence: it compares models and coding harnesses, describes a custom MCP-based setup, shows benchmark-style results, and links to supporting data and GitHub. It offers enough substance for a useful directory listing.