Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.10et.ai/llms.txt

Use this file to discover all available pages before exploring further.

Skills are instruction files that help AI agents do specific tasks better. They live in .claude/skills/ and get loaded when a task matches their trigger.

How Skills Work

Agent gets task → reads SKILL.md → follows instructions → produces output

Journal captures what worked/didn't → skill-learner harvests learnings

Behavioral eval measures: did the skill help? → build agent improves SKILL.md

Better skill → better outcomes → loop compounds

Quick Start

# See what skills you have
tenet skills list

# Check skill health (usage stats + learnings)
tenet skills status

# Install a skill from the registry
tenet skills install pretext

# Harvest learnings from your journal
tenet skills harvest brand-architect

# Improve a skill using behavioral eval
tenet skills improve pretext

Skill Levels

LevelLocationDescription
GlobalShips with @10et/cli24 skills everyone gets
Project.claude/skills/Specific to this project
Registry10et-ai/skillsCommunity skills, installable

Writing a Skill

Create .claude/skills/my-skill/SKILL.md:
---
name: my-skill
description: What this skill helps agents do
---

# My Skill

Instructions for the agent...

## When to Use
- trigger phrase 1
- trigger phrase 2

## How to Do It
1. Step one
2. Step two

## Anti-Patterns
- ❌ Don't do this
- ❌ Don't do that

## Learnings (auto-accumulated)
- Things learned from real usage get appended here

The Improvement Loop

Skills get better automatically through usage:
1

Agent uses skill

Session reads SKILL.md, follows instructions, produces output. skill-tracker.ts logs the read event.
2

Journal captures outcome

Session journal records what worked, what failed, what the human fixed.
3

Harvest learnings

tenet skills harvest <name> extracts learnings from journals. Patterns: fixes after skill use, decisions overriding defaults, discoveries.
4

Behavioral eval measures

Eval tests: does using this skill produce working output? NOT: does SKILL.md contain specific strings.
5

Build agent improves

tenet skills improve <name> runs a build agent that rewrites SKILL.md using accumulated learnings + behavioral eval as the reward function.

Behavioral Evals

Skills have evals that test outcomes, not implementation:
// ❌ BAD — tests what SKILL.md says
checks.push({ pass: skillMd.includes("layoutNextLine") })

// ✅ GOOD — tests what the skill produces
checks.push({ pass: outputFileExists && pageLoadsWithoutError })
Current skill evals:
SkillScoreChecks
pretext10/10API docs, examples, anti-patterns, learnings
content-creator7/7Hook, platform, voice, CTA, content types
brand-architect5/8Needs color/typography/quality guidance

Publishing Skills

Share your improved skill with the community:
tenet skills publish my-skill
This clones the registry, adds your skill, and creates a PR.

See Also