Skip to content
@AgentGameLab

虾球球XiaQiuQiu

Building AI-Agent-native games and the infrastructure behind them. Games where agents don't assist — they are the game.

Popular repositories Loading

  1. typeless-but-free typeless-but-free Public

    Local, private, hold-to-talk voice dictation with AI cleanup. Like Typeless — but free & open source.

    Python 7 2

  2. mneme mneme Public

    Persistent long-term memory for Claude Code agents — SQLite + FTS5 + sqlite-vec + RRF hybrid recall over MCP, with Chinese tokenization. Local-first, graceful degradation.

    JavaScript 7 1

  3. learn-coach learn-coach Public

    教学陪练 Claude Code skill — 把「给答案」升级为「教到会」(费曼回讲 / 苏格拉底提问 / 脚手架分层 / 生产性失败)。A Claude Code skill that teaches to mastery, not just answers.

  4. harness-evolve harness-evolve Public

    Self-evolving Agent Skill (SKILL.md) for any harness with config + logs — one pipeline: track frontier research, self-audit config, act within L1/L2/L3 risk tiers. Native in Claude Code; also usabl…

  5. tuneup tuneup Public

    Agent Skill (SKILL.md) that turns a fuzzy request into a high-information-density prompt — triage + 4-quadrant forging + domain-expert lens + theory anchors. Native in Claude Code; also usable in C…

  6. evolve evolve Public

    Turn your AI agent into a capability amplifier — it watches your real work, finds where you're actually stuck, traces it to the discipline that studied it for decades, and feeds what it learns into…

    JavaScript

Repositories

Showing 9 of 9 repositories
  • tuneup Public

    Agent Skill (SKILL.md) that turns a fuzzy request into a high-information-density prompt — triage + 4-quadrant forging + domain-expert lens + theory anchors. Native in Claude Code; also usable in Cursor / Cline / Codex CLI or any agent that reads a markdown file. 把模糊需求翻译成高信息密度 Prompt。

    AgentGameLab/tuneup's past year of commit activity
    0 MIT 0 0 0 Updated Sep 24, 2026
  • mneme Public

    Persistent long-term memory for Claude Code agents — SQLite + FTS5 + sqlite-vec + RRF hybrid recall over MCP, with Chinese tokenization. Local-first, graceful degradation.

    AgentGameLab/mneme's past year of commit activity
    JavaScript 7 MIT 1 0 0 Updated Sep 23, 2026
  • atom-sfx Public

    Offline foley pipeline: long take-recording to named, axis-annotated sound atoms, then composed into game-ready variants. Companion to AnchorSFX.

    AgentGameLab/atom-sfx's past year of commit activity
    JavaScript 0 MIT 0 0 0 Updated Sep 12, 2026
  • skills Public

    AgentGameLab's agent skills as one installable Claude Code plugin — tuneup (fuzzy request → high-density prompt), harness-evolve (self-evolving agent harness, L1/L2/L3 risk tiers), learn-coach (teach to mastery, not just answers). Works in any harness that reads SKILL.md.

    AgentGameLab/skills's past year of commit activity
    Python 0 MIT 0 0 0 Updated Sep 7, 2026
  • OpenKOS Public

    Team memory engine for AI agents — write, route, graph, recall, inject. Self-hostable.

    AgentGameLab/OpenKOS's past year of commit activity
    JavaScript 0 MIT 0 0 0 Updated Sep 3, 2026
  • typeless-but-free Public

    Local, private, hold-to-talk voice dictation with AI cleanup. Like Typeless — but free & open source.

    AgentGameLab/typeless-but-free's past year of commit activity
    Python 7 MIT 2 0 0 Updated Aug 17, 2026
  • learn-coach Public

    教学陪练 Claude Code skill — 把「给答案」升级为「教到会」(费曼回讲 / 苏格拉底提问 / 脚手架分层 / 生产性失败)。A Claude Code skill that teaches to mastery, not just answers.

    AgentGameLab/learn-coach's past year of commit activity
    0 MIT 0 0 0 Updated Aug 9, 2026
  • harness-evolve Public

    Self-evolving Agent Skill (SKILL.md) for any harness with config + logs — one pipeline: track frontier research, self-audit config, act within L1/L2/L3 risk tiers. Native in Claude Code; also usable in Cursor / Cline / Codex CLI / LangGraph. Agent 系统自进化引擎。

    AgentGameLab/harness-evolve's past year of commit activity
    0 MIT 0 0 0 Updated Aug 9, 2026
  • evolve Public

    Turn your AI agent into a capability amplifier — it watches your real work, finds where you're actually stuck, traces it to the discipline that studied it for decades, and feeds what it learns into your mneme. Pairs with AgentGameLab/mneme.

    AgentGameLab/evolve's past year of commit activity
    JavaScript 0 MIT 0 0 0 Updated Jul 16, 2026

People

This organization has no public members. You must be a member to see who’s a part of this organization.

Top languages

Loading…

Most used topics

Loading…