Create new joelclaw skills with the idiomatic process — repo-canonical, symlinked, git-tracked, slogged. Triggers on 'add a skill', 'create skill', 'new skill', 'canonical skill', 'make a skill for', or any request to formalize a process or domain into a reusable skill.
95
Does it follow best practices?
Validation for skill structure
SKILL.md frontmatter and section structure
Frontmatter block present
0%
100%
name field
0%
100%
displayName field
0%
100%
description field
0%
100%
version is 0.1.0
0%
100%
author is joel
0%
100%
tags field
0%
100%
Trigger phrases in description
0%
100%
When to Use section
0%
100%
Operations section
0%
100%
Rules section
0%
100%
Written for another agent
62%
100%
Without context: $0.3750 · 2m 4s · 20 turns · 80 in / 6,360 out tokens
With context: $0.4477 · 1m 51s · 23 turns · 185 in / 5,938 out tokens
Canonical directory and symlink deployment
Canonical repo path
100%
100%
Directory name matches name field
100%
100%
Symlink to pi agent skills
0%
100%
Symlink to agents skills
100%
100%
Symlink to claude skills
100%
100%
Uses ln -sf not cp
85%
100%
All three destinations covered
66%
100%
Source is repo directory
100%
100%
Without context: $0.4768 · 1m 54s · 25 turns · 31 in / 6,794 out tokens
With context: $0.4393 · 1m 34s · 23 turns · 29 in / 5,497 out tokens
Slog and git commit workflow
slog write subcommand
0%
100%
slog --action configure
0%
100%
slog --tool skills
0%
100%
slog --detail flag
0%
100%
slog --reason flag
0%
100%
Commit message prefix feat(skills)
0%
100%
Commit message add pattern
0%
100%
git add targets skill dir
83%
100%
One skill per concern
100%
100%
Without context: $0.2053 · 52s · 12 turns · 14 in / 2,936 out tokens
With context: $0.4137 · 1m 39s · 21 turns · 26 in / 5,357 out tokens
Table of Contents
If you maintain this skill, you can claim it as your own. Once claimed, you can manage eval scenarios, bundle related skills, attach documentation or rules, and ensure cross-agent compatibility.