fix: ensure litellm is installed for LLM backend, .env never overrides HuggingFace secrets, add /ping endpoint

fix: add litellm openai/ prefix for NVIDIA model, inject provider env vars, robust LLM init

fix: strip vendor-specific LLM kwargs; expose per-backend errors

fix: remove litellm pin (conflicts with crewai==1.9.3's openai dep)

debug: expose LLM error in /health; pin crewai==1.9.3 + litellm==1.81.8

fix: add Groq as LLM fallback (NVIDIA → Groq → Local)

fix: point Vercel frontend to HF Space backend

ci: retrigger deploy with HF_TOKEN secret now set

fix: revert to repo-level HF_TOKEN secret (drop environment binding)

fix: use HF_TOKEN environment + secrets.HF for deploy workflow

feat: rewarding build UX + Docker fixes + web docs cleanup

feat: add expandable artifact file list to HITL approval cards + download endpoint

fix: remove duplicate ChipSummary declaration causing Vite error

feat: premium Claude-style UI for HITL approval cards and autonomous result screen

fix: sync LLM API secrets to HF Space on every deploy

feat: Supabase auth + Razorpay billing + fix API double-URL bug

fix: frontend distinguishes backend-offline vs LLM-key-missing errors

fix: CI smoke test — add tests/__init__.py, fix discovery pattern to test_*.py

fix: use correct HF username vxkyyy in deploy.yml auth URL

fix: deploy.yml target correct HF space vxkyyy/AgentIC

feat: connect frontend to HuggingFace Space backend

docs: comprehensive deployment guide, HITL flow, CI/CD explanation, updated README

feat: add Docker + HuggingFace Spaces deployment pipeline

feat: implement contracts and enhance UI descriptions for clarity and accuracy

feat: add skip coverage option to Design Studio and Human In Loop Build pages

feat: Introduce a new comprehensive benchmarking suite with metric collection and reports, and remove obsolete design files.

docs: rewrite README — IP-safe, brand-matched, world-class

docs: update README with core module status, session 7-8 fixes, and clean architecture

Fix: multi-file Verilator/sby, width parser, TB stimulus storage, surgical RTL fixer, signal-level analyst, TB regen after RTL fix

feat: add HITL webapp with 3 UX improvements

feat(core): integrate core modules, update UI and docs

chore: re-add files to gitignore

chore: stop tracking results and training files

chore: track result and training files

chore: security audit, fix API leak, and update gitignore

fix: replace native EventSource with fetchEventSource to bypass ngrok block

fix: Add ngrok bypass header and safe fallbacks to prevent white screen crashes

fix: Update HTML page title to AgentIC

fix: Remove unused vars to fix Vercel strict TS build

feat: universal chip support, claude theme, and SFT logging export

Tier-1 upgrade: fail-closed orchestration, CI, and benchmark export

Update README with VeriReason benchmarks and functioning details

Update LLM config to Nemotron/GLM5, fix orchestrator write errors, update temp to 0.5

Update modified files and add LICENSE

Update AgentIC: Verification standards, Documentation, and Strict LLM Policy

feat: autonomous self-healing pipeline + pre-synthesis validation

Added interactive chat command with Groq and exposed VLSI tools

Implement Gate-Level Simulation (GLS) tool and bridge the logic-physical gap

Fix Mermaid syntax in README

Update source files and scripts