Fifteen reusable agent skills for curating LLM context windows, treating attention decay—not token capacity—as the real constraint. A routing layer benchmarked at 0.92 top-1 accuracy selects the right skill for each task.
AI Client2025
Run official Claude Code/Codex/Gemini/OpenCode coding sessions locally and control them remotely via Web/PWA/Telegram — local-first workflow with seamless handoff between local and remote, AFK approvals from phone, voice assistant, and multi-agent CLI support.
AI Deploy2026
Enables parallel speculative decoding by using a lightweight block-diffusion draft model to produce multi-token drafts for faster, high-quality generation. Integrates with vLLM, SGLang and Transformers backends and ships draft models on Hugging Face.
AI Client2026
Provides a Python API, CLI, and agent skill to programmatically access Google NotebookLM — exposing features the web UI omits (batch imports/exports, PPTX slides, mind‑map JSON) and integrating with LLM agents like Claude Code and Codex.
AI Audio2026
Real‑time full‑duplex speech‑to‑speech system that controls conversational role via text prompts and voice timbre via audio-conditioned embeddings. Built on Moshi; optimized for low-latency, persona-consistent spoken interactions.
AI Client2026
Transforms unstructured documents into strongly-typed Knowledge Abstracts with one CLI command, extracting entities and relations into graphs, hypergraphs, and spatio‑temporal structures. Includes 80+ templates, multiple RAG engines, local vLLM support, Obsidian export and an MCP server.
MCP Client2026
Compresses any context sent to LLMs (tool outputs, DB reads, RAG results, files, logs) to cut tokens by ~70–95% while preserving reversible originals; runs as a proxy or Python/TypeScript SDK with integrations for common agent frameworks.
AI Client2026
Aggregates global news, infrastructure, military and market signals into an interactive map dashboard and synthesizes AI-generated intelligence briefs. Key features: local/remote LLM support, 3D globe + flat map, 35+ data layers, country instability index and client-side RAG/embeddings.
AI Client2026
Generates daily LLM-powered decision dashboards for A/H/US stocks by combining multi-source market data, real-time news, technical signals and agent-style strategy reasoning; deploys via GitHub Actions or Docker and pushes reports to multiple channels.
AI Agent2026
Teaches LLMs to detect and remove “AI tells” from prose using curated phrase/structure lists, before/after examples, and a 5‑dimension scoring rubric. Delivered as a reusable skill (SKILL.md + reference files) designed to plug into Claude or any LLM workflow for automated style sanitization.
Provides a conditional memory module that performs O(1) N‑gram lookups and fuses static embeddings into transformer hidden states — enables offloading large embedding tables to host memory with minimal inference overhead.
AI Agent2026
Grows a personal skill tree by crystallizing each solved task into reusable skills; a ~3K-line autonomous agent framework that gives an LLM system-level control of browser, terminal, filesystem, input/vision and mobile (ADB) via nine atomic tools, optimized for low token cost.