chore: trigger Space rebuild to load new Gemini API key secret

docs: update README with honest benchmark status (eval in progress)

fix: use treeless clone (--filter=blob:none) so old base_commits are reachable

fix: pass base_commit to API so patches apply at correct SWE-bench commit

fix: add Gemini provider block in reflection_agent.py (httpx REST, no SDK)

fix: GeminiClient now uses httpx REST (no deprecated google-generativeai SDK)

ci: fix workflow — use lightweight requirements-test.txt (244 tests pass)

ci: add GitHub Actions workflow (244 tests) + CI/status badges in README

docs: add complete improvement roadmap for top-tier AIML resume

fix: exclude task_id from **result kwargs to avoid duplicate argument error

fix: strip trailing newline from GROQ_API_KEY before using in HTTP header

fix: use httpx directly for Groq API (bypass SDK connection issues in HF Spaces)

fix: route LLM to Groq (deepseek-r1) instead of hardcoded gpt-4o/openai

fix: skip git checkout when base_commit is empty/HEAD — fixes clone error in HF Spaces

fix: add HuggingFace Spaces config to README

chore: remove .venv and node_modules from git tracking

fix: use slim requirements-api.txt in Dockerfile, exclude torch/GPU packages

feat: add Dockerfile for HuggingFace Spaces deployment

fix: auto-load .env in llm_client; add SWEBenchLoader class to loader

fix: normalise LangGraph dict output back to AgentState in _run_with_langgraph

docs: add complete project guide (setup, learning roadmap, deployment, interview prep)

Initial commit