CentificAIResearch/sales-revops-agent-rl-env

Update app.py (#1)

feat(eval): add Tool Calls + Think Steps chips to Live Run score strip

feat(eval): match Live Run eval to the dark Load-Run report style

fix: filter Gemini heatmap to 7 comparable rubrics (drop 2 Gemini-only extras)

fix: restore Gemini reward signal matrix (remove allowed_rubrics filter)

preload revops_10 report on page load instead of showing placeholder

update revops_10 report: multi-model (Gemini/Sonnet/GPT-5.5) with failure mode analysis

fix(report): serve report as static file, not a 1.6MB base64 data: URI

Upload cached_runs.json with huggingface_hub

Delete files reports/revops_10_meta_v2.html with huggingface_hub

Upload cached_runs.json with huggingface_hub

Upload app.py with huggingface_hub

Upload reports/revops_10.html with huggingface_hub

Upload app.py with huggingface_hub

Upload reports/revops_10_meta_v2.html with huggingface_hub

Anonymized single-task release: revops_10 only (cached, no OpenRouter)