Upload clean calibrated runtime

Clean Space runtime files

Calibrate agent search recommendations

Add reflective agent loop and hidden gem evidence

Deepen dataset search and refine agent trace

Polish submission copy and candidate board naming

Improve candidate board readability and relevance checks

Replace candidate circles with evidence board

Remove ranked-result selected highlight

Redesign the agent research workspace

Improve task-aware dataset relevance and use Sora

Refresh interface with Codex light blue theme

Fix schema-aware dataset ranking

Run SmolLM2-360M locally for agent planning

Deploy HF Agentic Search for Build Small

initial commit