>_Skillful
Need help with advanced AI agent engineering?Contact FirmAdapt
ranxianglei

ranxianglei

@ranxianglei

On GitHub since May 2015

18

Published Tools

0

Total Stars

1,781

Weekly Downloads

44

GitHub Followers

44

Public Repos

Published Tools

1 MCP Server13 Skills4 Agentsacross 5 categories

Context-compression proxy that lets AI coding agents run for days — billions of tokens through one context window. Sits between any agent (Claude Code, Codex, Cursor, Aider, …) and its model API, folding consumed conversation into reversible, prefix-cache

...more
SkillLLM Tool
1 dir

One billion, not one million. Model-driven context management for claude-code. Placeholder — real release coming soon. See billion-context-pi for the working implementation.

...more
MCP Servermcp
1 dir

Node sqlite storage backend for pi-stable-agent-core sessions

Skillai-ml
17/wk1 dir

General-purpose agent with transport abstraction, state management, and attachment support

AgentLLM Tool
1 dir

ranxianglei

Active Context Pruning — model-driven context management for OpenCode (hardened fork of DCP with 34 bug fixes)

SkillAI Tool
1.8K/wk1 dir

ranxianglei

ework-web — standalone multi-project issue tracker. Local SQLite-backed, no external API dependency. Bun + TypeScript + SSR HTML.

...more
SkillAI Tool
1 dir

ranxianglei

Issue-driven AI development daemon. Spawns opencode subprocesses to resolve Gitea issues.

AgentAI Automation
1 dir

ranxianglei

Standalone webhook router for ework — receives events, routes to daemons by strategy, proxies replies

SkillAI Tool
1 dir

ranxianglei

Active Context Pruning (ACP) extension for Pi — model-driven context compression that keeps conversations flowing.

SkillLLM Tool
1 dir

Standalone algorithms for conversation context compression — quality gates, prompt providers, trigger policies. MIT-licensed.

...more
SkillLLM Tool
1 dir

ranxianglei

All-in-one installer for ework (issue tracker) + ework-daemon (AI bridge) + opencode-ework (plugin). One command: npm i -g ework-aio && ework-aio install.

...more
AgentAI Tool
1 dir

One billion, not one million. Model-driven context management for the Pi coding agent.

SkillLLM Tool
1 dir

ranxianglei

Universal context-compression proxy for AI coding agents. Sits between any agent and its model API, rewriting Anthropic/OpenAI streams with acp-kernel compression. Any agent that can set a base URL (Claude Code, Codex, Cursor, Aider) works out of the box.

...more
SkillLLM Tool
1 dir

ranxianglei

Unified LLM API with automatic model discovery and provider configuration

AgentLLM Tool
1 dir

One billion, not one million. Model-driven context management for the oh-my-pi (omp) coding agent.

SkillLLM Tool
1 dir

ranxianglei

Framework-agnostic context-compression engine (model-driven, 3-tier LSM). Pure core: no host dependency.

SkillLLM Tool
1 dir

Host-agnostic shared surface for billion-context adapters (panel rendering, range viability, topic fallback). Layer 3 of the acp-kernel → kit → adapter stack.

...more
SkillAI Tool
1 dir

ranxianglei

One-way mirror from ework to Gitea (webhook-driven, Gitea-compat auth as awork user)

SkillAI Tool
1 dir