scriptifylisted
Install: claude install-skill hacxy/skills
# Scriptify
Evaluate a skill's steps for scriptability, then extract all scriptable parts into executable scripts. For skills that have already been scriptified, audit existing scripts for quality and coverage gaps.
**Core principle:** If a script can solve it, don't use a skill for it. The LLM should only handle tasks that require judgment, creativity, or error recovery.
## Mode Detection
Before starting, check whether the target skill already has a `scripts/` directory with executable scripts:
- **Has scripts → Audit Mode** (Phase A): review existing scripts for rigor, then check for new scriptification opportunities
- **No scripts → Scriptify Mode** (Phase 1–4): the original assess-and-convert workflow
The user can also explicitly request audit mode ("audit scripts", "check scriptification", "review my scripts").
## Report Language
All reports and user-facing output should match the user's language. Detect the language from the user's message — if they write in Chinese, output the report in Chinese; if in English, output in English; and so on for other languages. The report templates below are shown in English as examples, but translate all headings, labels, severity names, and summary text to match the user's language when presenting.
## Script Language Priority
Detect the user's OS and choose the script language accordingly:
| OS | Priority |
|----|----------|
| macOS / Linux | shell > js > python |
| Windows | PowerShell > js > python |
Check OS:
```bash
una