本文へ移動
cccskills
無料GitHub で公開

mini-openamer

Use when the agent needs recursive reasoning, episodic memory retrieval, internet learning, or world-model prediction. Mini-OpenAmer: the 2B hybrid Mamba core with 9 tools, running 24/7 on Damir's laptop.

インストール方法を見る

含まれるファイル(1)

  • SKILL.md2.6 KB

SKILL.md(原文)

インストールする前に、エージェントに与えられる指示の中身を確認できます。

Mini-OpenAmer Integration

Mini-OpenAmer is the self-learning core that runs alongside the main OpenAmer agent. It provides capabilities the main agent can call:

What it provides

  • Recursive Reasoning: think → critique → improve loop
  • Episodic Memory: 3.012+ episodes with 768d embeddings
  • Internet Learning: 5 source types, 24/7 (news, papers, GitHub, docs, competitors)
  • World-Model: cause→effect graph with future predictions
  • Meta-Learning: adapts its own learning rate and strategy
  • Self-Improvement: modifies its own code with test gates
  • Swarm Intelligence: laptop + PC as collective (task routing, consensus)
  • Auto-Skill Creation: internet insights become new Darwin skills

Integration points

ComponentLocationPurpose
Tool Server:80819 tools + test-time training
Smart Routersmart_router.py3-tier routing (local→cloud→GPU)
Frontier ServerPC :8082Qwen3.5-4B deep reasoning
Darwindarwin_engine.pySkill evolution, 15 min cycles
Self-Modelmemory/self_model/identity.mdEvolving identity
World-Modelmemory/world_model.jsonlCause-effect + predictions
Meta-Statetraining/meta_state.jsonLearning-process self-knowledge

Energy

  • Laptop: ~25 W (2B + all learning loops)
  • PC (GPU worker): ~35-50 W (4B warm, training on demand)
  • Total: ~60-75 W → target 20 W via quantization/distillation/Akida
  • Running cost: 0 € (no API fees, local hardware)

レビュー

まだレビューはありません。使ってみた感想をお寄せください。

同じリポジトリのスキル

概要と使いどころ

Build integrated IS/BS/CF financial workbooks in Excel.

日本語の概要は準備中です。原文の説明を表示しています。

openamer/openamer62026年10月12日 更新

a2a-swarm

無料

Run and grow the OpenAmer Agent-to-Agent swarm: identity, trust, node-to-node ask, signed skill/insight sharing, and the autonomous self-learning loop.

日本語の概要は準備中です。原文の説明を表示しています。

openamer/openamer62026年10月12日 更新

Use for A/B experiments on OpenAmer configs and skills.

日本語の概要は準備中です。原文の説明を表示しています。

openamer/openamer62026年10月12日 更新

Roleplay a hostile user to find and triage UX pain points.

日本語の概要は準備中です。原文の説明を表示しています。

openamer/openamer62026年10月12日 更新

Agent mesh: master/worker nodes, HTTP delegation, heartbeat.

日本語の概要は準備中です。原文の説明を表示しています。

openamer/openamer62026年10月12日 更新

Use for agentic commerce payments tasks. Candidate synthesized from trend radar; awaiting trial promotion.

日本語の概要は準備中です。原文の説明を表示しています。

openamer/openamer62026年10月12日 更新

openamer のスキルをすべて見る

このスキルの問題を報告する