feat: enforce strict CORS and SlowAPI rate limiting

feat: add startup-grade backend routes for massive credit refills, SaaS policy settings, and real-time logs fetch

feat: implement POST /me/share API endpoint to safely reward users with +500 K2 Credits

feat: implement secured admin route router, endpoints, user is_admin property helper, and auto-seeding promo script

style: enhance Matplotlib chart with premium dark background, value labels, and tight margins

fix: return FileResponse for chart POST endpoint instead of JSON

fix: indent JSON parsing block inside for loop

fix: merge multiple markdown json blocks to prevent dropping keys

fix: handle lists returned by K2 for string fields in ExperimentalProtocol

fix: robust extraction of nested JSON keys from K2 engine output

fix: static chart serving and JSON URL response for Strategy Visualizer

fix: optimize K2 engine performance with reduced tokens and conciseness to prevent timeouts

fix: gracefully catch ArXiv HTTP 429 errors and fix export analysis endpoint parameter binding

feat: enhance protocol schema and add audit log traceability

fix: prioritize full block JSON extraction and fix comparative_analysis search

chore: update K2 JSON schema order, update extraction strategy, and add debugging files

fix: harden JSON parsing and handle missing think tags

Fix Pydantic validation errors for fallback objects

Bulletproof JSON extraction, strict prompts, and salvaged text fallback

Fix f-string syntax error: escape curly braces

Fix syntax error: restore missing triple quotes in K2 prompt

Enhance LLM prompt and regex extraction for JSON robustness

Fix chat UUID mapping and harden JSON auto-repair

Fix Pydantic ValidationError for ExperimentalProtocol missing fields

Fix K2 timeout detection and JSON parsing errors

Implement adaptive retry for K2 timeouts

Fix schema nesting for frontend compatibility

Total rewrite of K2ThinkEngine for robustness

Optimize K2 prompt and add JSON repair logic

Fix NameError, improve logging and UUID handling

Capture errors and stabilize LLM output

Hardened JSON extraction and normalization logic

Fix K2 Think JSON extraction by stripping <think> tags

Fix analysis normalization logic and missing fields

Fix IndentationError in analysis.py

Fix API route conflict causing 404 on /chat endpoints and optimize LLM params

Add final validation to ensure displayable content

Add result normalization to ensure frontend compatibility

Add debug logging to see K2 result structure

Fix analysis retrieval to work with mismatched project_ids from frontend

Force K2 Think real analysis mode by default instead of mock data

Fix K2 Think JSON parsing and secure default configs

Fix robust JSON regex extraction to bypass unclosed think blocks

Fix K2 Think JSON parsing and infinite loop

Fix frontend data mismatch by flattening result_data in GET analysis response

Aggressively enforce complete JSON generation in system prompt to prevent truncation

Fix Pydantic validation errors by coercing strings to lists in Protocol parser

Fix Matplotlib backend issue for Strategy Map visualizer

Remove max_tokens limit and strongly discourage long thinking block

Limit K2 reasoning length to prevent JSON truncation