Trustbaselab

Trustbase Skills — AI skill for Claude Code

AI community

Agent Skills for AI that connect to verifiable facts — 8 cognitive skills (built from classic books) + 6 industry skills (rCB / CBAM / compute PUE).

How to install Trustbase Skills

This entry records only its repository, not the path inside it, so there is no exact command to give. Open Trustbaselab/trustbase-skills and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Trustbase Skills does

Agent Skills for AI that connect to verifiable facts — 8 cognitive skills (built from classic books) + 6 industry skills (rCB / CBAM / compute PUE). Apache-2.0.

Alternatives in AI

Agent Rules Books — AGENTS.md rules / skills for AI coding agents: Codex, Cursor & Claude Code 2.6k ★
Fact Checker — Verify claims and facts with sources 653 ★
Investment News — 为 A股投资者打造的全球产业链资讯看板 · 12 大赛道一一对应 A股板块(半导体/AI/机器人/新能源车…)，覆盖 100+ 权威源，用你自己的大模型每日提炼中文「今日要点」+ 翻译 · 全程本地、零 API key 447 ★

README

Trustbase Skills

**8 skills that turn classic books into executable rules for AI agents.** **Not summaries — if a sentence can't become "when X, do Y", it doesn't go in.**

[![skills.sh](https://skills.sh/b/Trustbaselab/trustbase-skills)](https://skills.sh/Trustbaselab/trustbase-skills) [![License: Apache-2.0](https://img.shields.io/badge/License-Apache_2.0-blue.svg)](LICENSE) [![Skills](https://img.shields.io/badge/skills-14-green.svg)](#skills) [![Verifiable facts](https://img.shields.io/badge/verifiable_facts-520+-orange.svg)](https://trustbaselab.com/database)

npx skills add Trustbaselab/trustbase-skills

**Quick nav**: [What is this](#what-is-this) · [The 8 skills](#the-8-skills) · [Why different](#why-this-is-different) · [Install](#install) · [Feedback](#feedback)

![The eight-layer cognition operating system](assets/cognition-os.svg)

What is this

**Cognitive skills** — rules an AI agent can actually execute, derived from classic books.

Each one was validated in a real project. The clearest example: applying *Scarcity* + *Deep Work* to our own task queue cut per-round context from **~141K tokens to ~5.6K (−96%)**.

There is also a meta-skill, `book-to-skill`, which packages **the method of turning a book into a skill** — including a scoring rubric for book selection, a seven-step SOP, and a delivery checklist.

**The design rule everything follows:** if a sentence from the book cannot be translated into "when X, do Y", it does not enter the skill. That rule filtered out about 60% of what we initially wanted to include.

The 8 skills

Skill	Answers the question
`factfulness`	Is this number trustworthy?
`thinking-fast-and-slow`	Is this AI conclusion reliable — or just fluent?
`noise-in-judgment`	Why does the answer change every time?
`thinking-in-systems`	Why does fixing it make it worse?
`bandwidth-and-deep-work`	Why can't I focus, and what do I do first?
`kn