skillscripts

Fast script search and skill retrieval CLI.

What It Is

skillscripts (alias sks) is a local-first command-line tool with two core capabilities:

Fast Script Search:

• You have scripts scattered everywhere and can't find them
• You want to quickly locate a script by functionality
• You need a lightweight script manager

Fast Skill Retrieval:

• You are an AI Agent developer managing a skill library
• You want to quickly retrieve reusable skills
• You need to provide tool-calling capabilities for Agents

Core Features

Dual-Mode Search

The same command serves two scenarios:

Script Search Mode:

• Quickly locate script files by functionality
• Lightweight script management without complex databases
• Instant results with fuzzy matching

Skill Retrieval Mode:

• Retrieve reusable skills for AI Agents
• Output YAML ready for tool calling
• Manage skill libraries for Agent development

Both modes share the same output format (YAML with name, description, tags, path), making it easy to use scripts as Agent tools.

Instant Scanning

• Parallel file scanning, millisecond response
• In-memory caching for repeated operations
• Automatic encoding detection, skip binary files
• gitignore rules support

Smart Matching

• Fuzzy matching on name, description, tags
• Intelligent scoring and sorting
• Paths excluded from search to reduce noise

YAML Header

Any script becomes searchable by adding a YAML header:

# ---
# name: resize_image
# description: Resize image using PIL
# tags: [image, python]
# ---
from PIL import Image

This header serves dual purposes:

• For script search: Provides metadata for quick identification
• For skill retrieval: Defines tool interface for Agent invocation

Supported comment styles: #, //, --, %, /**, etc.

Interactive Selection

skim-based TUI interface with live preview, suitable for both script selection and skill browsing.

Installation

From release:

• https://github.com/gtiders/skillscripts/releases/latest

From source:

cargo install --path .

Quick Start

# Initialize config
sks init

# Search scripts/skills
sks search image

# List all scripts/skills
sks list

# Interactive selection
sks pick

Commands

Command	Description
init	Create config file. Use --local for project-level config.
config	View current configuration.
search <query>	Fuzzy search, output YAML.
list	List all scripts/skills, output YAML.
pick	Interactive TUI selector.

Output Format

search and list output YAML:

- name: resize_image
  tags:
    - image
    - python
  description: Resize image using PIL
  path: ./scripts/resize_image.py

Configuration

Config file locations:

• Global: ~/.config/skillscripts/skillscripts.yaml
• Local: ./skillscripts.yaml (project-level, merged with global)

Configuration Example

scan_paths:
  - skills
  - ./scripts
  - ~/projects/utils
ignore_patterns:
  - target
  - .git
  - node_modules
max_file_size: 1MB
search_limit: 10
report_parse_errors: true

Configuration Options

Option	Description	Default
scan_paths	Paths to scan	["."]
ignore_patterns	Patterns to ignore	[]
max_file_size	Maximum file size	1MB
search_limit	Search result limit	5
report_parse_errors	Report parse errors	false
copy_to_clipboard_on_pick	Copy path to clipboard after pick	false

YAML Header Specification

Required Fields

Field	Description
name	Script/skill name
description	Script/skill description

Optional Fields

Field	Description
tags	Tag list
args	Parameter definitions
version	Version number
command_template	Command template

Examples

Python:

# ---
# name: disk_check
# description: Check disk usage
# tags: [ops, monitoring]
# args:
#   path:
#     type: string
#     description: Target path
#     required: false
# ---
import shutil
shutil.disk_usage(path)

Shell:

#!/bin/bash
# ---
# name: git_log
# description: Show recent commits
# tags: [git, vcs]
# ---
git log --oneline -10

JavaScript:

// ---
// name: fetch_data
// description: Fetch remote data
// tags: [http, async]
// ---
const response = await fetch(url);

License

MIT