jhammant

ClaudeHistoryMCP. MCP server for searching and surfacing Claude Code conversation history

Turbo1bit. Turbo1Bit: Combining 1-bit LLM weights (Bonsai) with TurboQuant KV cache compression for maximum inference efficiency. 4.2x KV cache compression + 16x weight compression = ~10x total memory reduction.

ilya-top-30. AI-powered learning platform for Ilya Sutskever's legendary AI paper reading list. Master the foundations of modern AI with curated paths, AI tutoring, and gamification.

11plus-tutor. Free, open-source 11+ exam practice - 1,364 verified questions for grammar school preparation

precog. 🔮 Sub-500ms voice agent with speculative pre-generation. The more you talk, the faster it gets.

agent-drift. Stress-test AI agents for goal drift and system prompt violations. Inspired by 'Asymmetric Goal Drift in Coding Agents' (ICLR 2026).

openclaw-squad. AI agent team setup with OpenClaw - 7 specialized agents working together

lobsterpot. 🦞 Remote coding agent orchestrator — manage Claude Code, Codex, Aider sessions via SSH+tmux

bad-reachy. Python

wanderlust. Discover your travel history from Apple Photos and get AI-powered holiday recommendations

claude-history-cloud. TypeScript

doac-insights. What 400+ world-class performers agree on — AI-powered analysis of every Diary of a CEO episode. Interactive knowledge graph, cross-episode patterns, and key learnings.

claude-history-mcp. TypeScript

self-healing-router. Graph-based self-healing tool routing for LLM agents. 93% fewer LLM calls than ReAct.

dyson-mechanic. Python

mcp-discord-bridge. MCP server that gives Claude Code native Discord tools — bridge to remote AI agents

hash_bench. Python

agent-dreams. AI agent dreaming system — replay, consolidation, creative synthesis during quiet hours

expertflow. ⚡ Dynamic MoE expert streaming for Apple Silicon — run 685B parameter models on a laptop. Rust-native GPU + ANE + NVMe orchestration.

nano-llm. Python

ESTAR-JOLO. Python

x86-perf-analyzer. Profile and optimize hot loops on x86 (Intel DSB + AMD OpCache). Find decode bottlenecks, uOP cache misses, branch mispredictions.

arm-perf-analyzer. Profile and optimize hot loops on ARM (Neoverse-N1 / Apple Silicon). Finds decode bottlenecks, cache misses, branch mispredictions.

Cloud-Gym. Benchmark and training data generation for LLM-based Infrastructure-as-Code repair via environment inversion

SecScan. Local-LLM-powered security scanner. Multi-lens per-file review + architecture extraction + threat-model synthesis + sandboxed exploit PoCs. 100% local inference via LM Studio.

worldclone. LLM forecasting toolkit: Halawi retrieval + multi-agent simulation + Kelly accumulator. Geopolitics, box office, sports. 100% local on M5 Mac.