Validated agent skills distilled from real-world practice.
This repository follows the open SKILL.md directory format used across agent ecosystems, including Codex, Claude Code, skills.sh, SkillsMP, and ClawHub/OpenClaw.
| Skill | Description |
|---|---|
| article-to-insights | Turn one article URL into a structured reading result: core summary, relevance analysis, and actionable recommendation. |
Install directly from GitHub with an explicit skill name:
npx skills add https://github.com/lesterli/braid --skill article-to-insightsThis repo currently contains one skill, but using --skill article-to-insights keeps the command stable if more skills are added later.
Prerequisite: the corresponding agent CLI is already installed.
# Clone the repo
git clone https://github.com/lesterli/braid.git
cd braid
# Claude Code global skills (~/.claude/skills)
mkdir -p ~/.claude/skills
cp -r article-to-insights ~/.claude/skills/
# Codex global skills (${CODEX_HOME:-~/.codex}/skills)
mkdir -p "${CODEX_HOME:-$HOME/.codex}/skills"
cp -r article-to-insights "${CODEX_HOME:-$HOME/.codex}/skills/"Restart the agent after installation so new skills are discovered.
skills.sh: install from this GitHub repo withnpx skills add ... --skill article-to-insights- SkillsMP: compatible with the open
SKILL.mdspec and suitable for indexing or listing as-is - ClawHub / OpenClaw: each skill directory is publishable as a standalone skill bundle
Each skill is a self-contained folder:
skill-name/
SKILL.md # Entry point: trigger rules, workflow, output format
references/ # Supporting docs: examples, guardrails, format specs
scripts/ # Helper scripts for validation or preprocessing
agents/ # Agent-specific configs (e.g., OpenAI, Claude)
MIT