which-tool_skill

This skill helps you pick the fastest, most suitable CLI tool for a task, with graceful fallback and proactive research when needed.
  • TypeScript

25

GitHub Stars

1

Bundled Files

2 months ago

Catalog Refreshed

4 months ago

First Indexed

Readme & install

Copy the install command, review bundled files from the catalogue, and read any extended description pulled from the listing source.

Installation

Preview and clipboard use veilstrat where the catalogue uses aiagentskills.

npx veilstrat add skill outfitter-dev/agents --skill which-tool

  • SKILL.md6.0 KB

Overview

This skill helps choose the best CLI tool for a given task, provide graceful fallbacks, and trigger research when tools underperform. It focuses on file search, content search, JSON processing, and other common developer CLI needs. The goal is faster, more ergonomic commands while never blocking the workflow.

How this skill works

On first use in a session it runs a detection script to list available modern tools, missing candidates, and system context (OS, package managers). It maps the task to a preferred tool using a curated selection table, uses fallbacks or legacy commands when needed, and surfaces installation suggestions or research prompts when gains are significant. Results are cached per session to avoid repeated detection runs.

When to use it

  • Choosing which CLI tool for file search, content search, JSON processing, or HTTP requests
  • When a tool feels unexpectedly slow for the task size
  • If you express frustration with the current tool or workflow
  • When a task could be done more elegantly with a specialized CLI
  • Before recommending a tool, to verify availability on the system

Best practices

  • Always run detection before recommending a specific tool; cache results per session
  • Prefer modern tools (fd, rg, jq, bat, delta, etc.) with optimal flags
  • Provide a non-blocking fallback and surface install suggestions for significant improvements
  • Trigger research when a tool is 3x+ slower than expected or lacks critical features
  • Avoid recommending unmaintained tools and never assume an unverified tool is installed

Example use cases

  • Find files by name: prefer fd, fallback to find; fd -e yaml -e yml
  • Search file contents: prefer rg (respects .gitignore) — e.g. rg "authentication" --type ts --type js
  • Process JSON: prefer jq; if missing, use node/python with a suggestion to install jq
  • View files with syntax highlighting: prefer bat instead of cat
  • Navigate directories quickly: prefer zoxide for frecency-based jumping

FAQ

A session detection script is executed to report installed tools and system context; results are cached for the session.

What happens if the preferred tool is unavailable?

Minor gains use the next-best option silently. Significant gains surface an install suggestion. Critical gaps offer choices: install now, proceed, or defer.

When is research triggered?

Research is triggered if performance is 3x+ worse than expected, the task needs a specialized tool, or you request better approaches.

Built by
VeilStrat
AI signals for GTM teams
© 2026 VeilStrat. All rights reserved.All systems operational