Demo: sync v0.1.27 (GPT-5 Responses API retry fix) + switch to openai/gpt-5.5

Demo: sync agent layer + skills from main (v0.1.22..v0.1.26)

Demo: revert model to poe/Claude-Sonnet-4.6

Anthropic backend: omit temperature for opus-4-7 (deprecated)

Demo: switch model to anthropic/claude-opus-4-7

Demo: show loading overlay immediately on Reset Demo

Demo: install osqp backend for qpsolvers (RCTD dependency)

Fix repeated clarification submit buttons not working

Demo: surface preloaded reference path + column in welcome modal

Demo: gray out Notebook button instead of hiding it

Demo: disable GPU-only skills, hide Notebook btn, note compute limits

Demo: remove yellow banner that broke 100vh layout

Merge branch 'main' of https://huggingface.co/spaces/CyhVVVV/stat-agent-demo into demo

Add HF Spaces demo deployment (demo branch)

initial commit

v0.1.22: Fix YAML frontmatter quoting, fix 3D view plane sizing

Bump version to 0.1.21

Fix RCTD Reference() default logger parameter

v0.1.20: Fix prompt logger token tracking, add pathway enrichment compare skill

v0.1.19: Logs modal, coord scale, square spots, 3D bounds, spot detection, scipy compat

v0.1.18: Improve init form with dataset hints and provider dropdown

v0.1.17: Add 14 new skills, expand optional dependencies

v0.1.16: Skill configuration panel, default_skill attribute, top_k=5

v0.1.15: Comprehensive chat restore audit — fix multi-step, exchange isolation

v0.1.14: Fix separate queries merging into one exchange in chat restore

v0.1.13: Fix multi-step clarification chains in chat history restore

v0.1.12: Fix original query capture and reply positioning in chat history

v0.1.11: Fix chat history restore for clarification/skill selection flows

v0.1.10: Session persistence, chat history restore, logout

Session persistence: form recall, reconnect on refresh, chat history restore, logout

v0.1.9: Fix nrrd→pynrrd dep, unblock pip install stat-agent[skills]

Fix nrrd→pynrrd dep name: unblocks pip install stat-agent[skills]

v0.1.8: Per-block output sections, add scikit-misc and igraph deps

v0.1.7: Show execution output, fix duplicate response, filter noisy output

v0.1.6: Bundle skills into package for pip install

v0.1.5: Show execution output, fix duplicate response, filter noisy output

v0.1.4: 3D viewer, markdown rendering, chat UX improvements

v0.1.3: Render markdown tables/lists in chat, fix execution issue detection

Don't flag truncated output as partial_success in analysis prompt

Only pass stdout to analysis LLM when __interpret__ is set

Render markdown tables/lists in chat, suppress benign execution issues

Bump to v0.1.2

Add complete skill dependencies to [skills] extra

v0.1.1: Bundle web UI into package, simplify deps

Add PyPI publish workflow and fix license format

Initial release v0.1.0