fix: Add robust error handling for all visualization builders

fix: Add critic evaluation to synthesis engine call

feat: Use actual trained agents from ForgeEngine + SynthesisEngine for real multi-perspective reasoning

feat: Showcase Codette's actual algorithmic reasoning — pure Python perspective generation + RC+xi synthesis

fix: Remove gr.Request() from inputs — Gradio auto-injects via function signature

feat: Implement OAuth-based auth flow with gr.LoginButton and request context

fix: Graceful fallback for perspective calls when auth token unavailable

feat: Replace manual token input with native Gradio LoginButton for OAuth

fix: Add graceful fallback synthesis when LLM API fails (401/403)

feat: Add HuggingFace token login panel to Gradio interface

fix: Gracefully handle invalid HF tokens without crashing startup

Merge remote changes and resolve HF_TOKEN authentication conflict

fix: Use huggingface_hub.login() for global token authentication

Update app.py

fix: Replace deprecated Plotly titlefont with title.font

fix: Handle spiderweb nodes as dictionary with state array instead of list

fix: Access memory.memories directly instead of memory.store method

fix: Access ResonanceState.psi_r attribute directly instead of using .get()

fix: Remove Gradio v6 incompatible parameters (type='messages' from Chatbot, move theme/css to launch)

fix: Remove invalid label parameter from gr.HTML visualization components

fix: Correct return tuple size for empty message case (15 values, not 12)

fix: Convert Plotly figures to HTML strings to avoid gr.Plot schema introspection bug in Gradio 5.x

feat: v2.0 - Add production Plotly visualizations for all 10 subsystems

Initial Codette cognitive architecture demo Space