Official agent skill

Agent Skill Finder

by google in google/adk-recipes

Searches skills.sh for agent skills you do not have yet, checks their quality, presents the option and installs only after you confirm.

OfficialApache-2.0Auto-check passedAgent Workflows

Install Agent Skill Finder

skills CLI
$ npx skills add google/adk-recipes --skill find-skills -a claude-code

Project install by default; add -g for ~/.claude/skills/.

GitHub CLI
$ gh skill install google/adk-recipes find-skills --agent claude-code

Project scope by default; add --scope user for a personal install. Needs GitHub CLI 2.90.0 or later (public preview).

Manual copy
$ git clone --depth 1 https://github.com/google/adk-recipes.git skills-src && mkdir -p .claude/skills && cp -r skills-src/core/python/long-horizon-harness/horizon/builtin_skills/find-skills .claude/skills/find-skills && rm -rf skills-src

Use ~/.claude/skills/ instead of .claude/skills for a personal install. The folder must contain SKILL.md.

Claude Code skills documentation · loads skills from .claude/skills/

Facts

Skill name
find-skills
GitHub stars
10k
Token cost
~1.1k tokens
SKILL.md length
492 words
Files
1
Skills in repo
14
Repo updated
First seen
Licence
Apache-2.0

At a glance

Searches skills.sh for agent skills you do not have yet, checks their quality, presents the option and installs only after you confirm.

  • Works in 6 steps: check the leaderboard → search for skills → verify quality before recommending → …
  • You ask whether a skill exists for a task, or want to add a capability
  • SKILL.md covers When to use this skill, Step 1 — check the leaderboard, Step 2 — search for skills and Step 3 — verify quality before…, plus 5 more sections
  • Calls npx

What it does

When you ask for a skill for some task, or how to do a common task that probably has one, the agent first checks the skills already installed in its context. If none fits, it looks at the skills.sh leaderboard and then searches with the Skills CLI through `npx`, falling back to browsing the site and pasting the SKILL.md if `npx` is missing. Search terms are short, such as react performance or pr review.

Search results are never enough to recommend a skill. The agent weighs install counts, the reputation of the publishing organization and the source repository's popularity, treating low numbers with skepticism. It then describes the candidate to you and waits for a yes. Installation either writes the SKILL.md directly into the project's skills folder and reloads the catalog, which always works, or uses the CLI, which needs Node.js.

When your agent uses it

  • You ask whether a skill exists for a task, or want to add a capability
  • Evaluating a third-party skill's source before installing it
  • Installing a skill from skills.sh into the current project

Example prompts

  • “Find a skill that helps with changelog writing.”
  • “Is there a skill for reviewing pull requests? Check quality before recommending one.”
  • “Install the React performance skill you found, now that I have confirmed it.”

Requirements

  • Network access to skills.sh
  • Node.js with npx for the CLI search and install route

Workflow steps

6 steps, taken from the step headings in SKILL.md.

  1. check the leaderboard
  2. search for skills
  3. verify quality before recommending
  4. present the option to the user
  5. install
  6. confirm activation

What it can do on your machine

Read from SKILL.md and the folder at commit c339821. It shows what the files ask for, not the result of running them.

  • Tool permissions

    Pre-approves nothing: there is no allowed-tools line, so your agent's usual permission prompts apply.

    From allowed-tools in the SKILL.md frontmatter.

  • Runs code

    Shell commands in SKILL.md call:

    • npx

    From the folder's file list and the shell code blocks in SKILL.md.

  • Network

    Links to these hosts (documentation or services it may open):

    • skills.sh

    From URLs in SKILL.md, links to its own repository left out.

  • Credentials

    Names no API keys, tokens, secrets or passwords.

    From names ending in _API_KEY, _TOKEN, _SECRET, _KEY or _PASSWORD in SKILL.md.

Context cost

Agent Skill Finder loads about 1.1k tokens when it runs. Until then it costs about 52 tokens; SKILL.md has 492 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~52
When it runs · the whole SKILL.md, loaded when a task matches
~1.1k

Estimates: characters ÷ 4, the usual rule of thumb; real counts depend on the model's tokenizer. Scripts and assets cost tokens only if the agent reads them.

Safety

Auto-check passed

The automated check found no risky patterns in SKILL.md.

Automated static check — not a guarantee. Review scripts before installing. It scans the text of SKILL.md for risky patterns (piping downloads into a shell, reading credential files, hidden Unicode, destructive commands); files beside SKILL.md are not scanned.

SKILL.md

The full file from google/adk-recipes at commit c339821, republished under its Apache-2.0 licence (© google). 492 words, ~1,078 tokens.

Download SKILL.mdSave it as .claude/skills/find-skills/SKILL.md (or your agent's skills folder).
name
find-skills
description
Discover/install NEW agent skills from external sources (skills.sh). Use for capabilities you lack, e.g. "find a skill for X". Do NOT use to list skills you already have; see `<available_skills>`.

Find skills

When to use this skill

  • The user asks "find a skill for X", "is there a skill that does X"
  • The user asks "how do I do X" where X is a common task with a likely existing skill
  • The user wants to extend Horizon's capabilities with a new skill

Before searching externally, check the <available_skills> block in your context — the user may already have a relevant skill installed.

Step 1 — check the leaderboard

Before running any CLI command, check skills.sh to see whether a well-known skill already covers the domain. The leaderboard ranks by total installs, surfacing the most popular and battle-tested options.

Step 2 — search for skills

If the leaderboard doesn't cover the user's need, search via the Skills CLI:

bash(command="npx --yes skills find <query>")

Examples:

User saysSearch
"make my React app faster"npx --yes skills find react performance
"help with PR reviews"npx --yes skills find pr review
"create a changelog"npx --yes skills find changelog

If npx is unavailable the command will fail — fall back to browsing skills.sh manually and pasting the SKILL.md content.

Step 3 — verify quality before recommending

Never recommend a skill based solely on search results.

  • Install count — prefer 1 000+ installs; treat anything under 100 with skepticism
  • Source reputation — official organisations (vercel-labs, anthropics, microsoft) are more trustworthy than unknown authors
  • GitHub stars — check the source repo; a skill from a repo with <100 stars is a yellow flag

Step 4 — present the option to the user

I found a skill that might help: "react-best-practices" from Vercel Engineering
provides React and Next.js performance guidelines (185K installs).

Shall I install it?

Only proceed to installation after the user confirms.

Step 5 — install

Option A — write SKILL.md directly (always works)

If you have the skill content (pasted by the user, retrieved from skills.sh, or constructed from the search result):

write(".agents/skills/<name>/SKILL.md", content)
load_skill(action="reload")

write creates the directory if it does not exist. load_skill(action="reload") refreshes the <available_skills> catalog (plus extensions and manifest) on the next turn.

Show full SKILL.md (191 more words)Show less
Option B — install via npx (requires Node.js)

First check whether npx is available:

bash(command="which npx")

If it is, install the skill. Select a single skill from a multi-skill repo with the @<skill> suffix (omit it to take the whole repo):

bash(command="npx --yes skills add <owner/repo>@<skill-name> -y")

The Skills CLI stages the skill (with its references/, assets/, and scripts/ subdirs) under .agents/skills/<skill-name>/ — exactly where Horizon loads skills from — so no move is needed. Just refresh:

load_skill(action="reload")

The staged directory name MUST match the name: field in the SKILL.md frontmatter — _load_skill_from_dir validates that and skips mismatched skills.

Step 6 — confirm activation

After load_skill(action="reload") returns (look for the new skill in loaded), tell the user:

The "<name>" skill is now installed. It will appear in your
<available_skills> catalog on the next turn.

When no skills are found

  1. Acknowledge that no existing skill was found
  2. Offer to help directly using general capabilities
  3. Offer to write a custom skill for them:
bash(command="npx --yes skills init my-skill-name")

Or skip the CLI and author one manually — write a SKILL.md following the policy or workspace skills as examples.

Common skill categories

CategorySearch terms
Web developmentreact, nextjs, typescript, css, tailwind
Testingtesting, jest, playwright, e2e
DevOpsdeploy, docker, kubernetes, ci-cd
Documentationdocs, readme, changelog, api-docs
Code qualityreview, lint, refactor, best-practices
Designui, ux, design-system, accessibility

© google, Apache-2.0. Rendered from Markdown: HTML in the file is shown as text, images as links, and headings moved down two levels. Raw file

Files

Just SKILL.md in core/python/long-horizon-harness/horizon/builtin_skills/find-skills of google/adk-recipes.

Open the folder on GitHubat commit c339821

Compare with similar skills

Agent Skill Finder next to the 5 skills that share the most tags, products or categories with it. Stars are the repository's; “used in” counts other GitHub owners with a copy.

Agent Skill Finder compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Agent Skill Finder this skillgoogle/adk-recipes10k—~1.1kAutomated safety check: PassApache-2.0
Skill Base CLIginuim/skill-base1201 repos~1.9kAutomated safety check: PassNone
TeamAI Team SyncTencent/teamai-cli5.1k—~632Automated safety check: PassCustom licence
Darwin Skill Optimizeralchaincyf/darwin-skill6.2k1 repos~4.7kAutomated safety check: PassMIT
Using Agent Skillsaddyosmani/agent-skills103k4 repos~2.4kAutomated safety check: PassMIT
Ponytail Help CardDietrichGebert/ponytail158k—~728Automated safety check: PassMIT

Similar skills

  • Skill Base CLI

    ginuim/skill-base

    Uses the skb command to search, install, update, delete, publish and import skills on a Skill Base site, including curated collections and GitHub imports.

    120 GitHub starsUsed in 1 repo~1.9k tokens
    Agent WorkflowsAuto-check passed
  • TeamAI Team Sync

    Tencent/teamai-cli

    Make every team AI native — TeamAI syncs a team's AI skills, rules, docs and env across AI coding tools. Use when the task operates on team-shared AI…

    5.1k GitHub stars~632 tokensUpdated yesterday
    Knowledge ManagementAuto-check passed
  • Darwin Skill Optimizer

    alchaincyf/darwin-skill

    Scores SKILL.md files on a nine-dimension rubric, then improves them in a keep-or-revert loop with independent judge agents, test prompts, git history and human checkpoints.

    6.2k GitHub starsUsed in 1 repo~4.7k tokens
    Agent WorkflowsAuto-check passed
  • Using Agent Skills

    addyosmani/agent-skills

    Meta-skill for choosing which workflow skill fits the task at hand, plus always-on habits: surface assumptions, stop on confusion, push back, keep it simple and stay in scope.

    103k GitHub starsUsed in 4 repos~2.4k tokens
    Agent WorkflowsAuto-check passed
  • Ponytail Help Card

    DietrichGebert/ponytail

    Shows a one-shot quick-reference card for the ponytail skills: intensity levels, the six commands, and how to turn it off, set a default mode and update.

    158k GitHub stars~728 tokensUpdated today
    Agent WorkflowsAuto-check passed
  • Skill Creator

    zhayujie/CowAgent

    Guides creating, installing and updating agent skills in a workspace: SKILL.md frontmatter, bundled scripts and references, with scripts to scaffold, validate and package.

    47k GitHub stars~4.7k tokensUpdated today
    Agent WorkflowsAuto-check: notes

More from google/adk-recipes

All 14 skills in this repo
  • Retail Product Search Agent

    google/adk-recipes

    Official

    Builds a retail product search agent on Google Cloud, from catalog ingestion into BigQuery and Vector Search to ADK scaffolding, evaluation and Cloud Run deployment.

    10k GitHub stars~3k tokensUpdated today
    Auto-check passed
  • Align Recipe pyproject.toml

    google/adk-recipes

    Official

    Brings a Python recipe's pyproject.toml in line with the repo's CI rules, either as a read-only dry run or by rewriting the file while keeping comments.

    10k GitHub stars~4.6k tokensUpdated today
    Auto-check passed
  • Official

    Generates a minimal tests/test_runnability.py for a Python agent recipe that imports the agent module and checks root_agent, adding only the mocks and env vars it needs.

    10k GitHub stars~3.3k tokensUpdated today
    Auto-check passed
  • Retail Virtual Try-On Agent

    google/adk-recipes

    Official

    Sets up a virtual try-on agent on Google Cloud that generates image and catwalk-video try-ons with Gemini, from first setup through local testing.

    10k GitHub stars~3.5k tokensUpdated today
    Auto-check passed
  • Scaffold Python ADK Recipe

    google/adk-recipes

    Official

    Creates a new Python recipe for the ADK recipes repository by running a scaffold script that copies template files, after confirming the output directory and recipe name.

    10k GitHub stars~931 tokensUpdated today
    Auto-check passed
  • Official

    Reviews a GitHub pull request and drafts a small set of inline comments in a human reviewing voice, each checkable from the line it points at, then posts them after approval.

    10k GitHub stars~8.4k tokensUpdated today
    Auto-check passed

Works with

Categories

Questions about Agent Skill Finder

What does Agent Skill Finder do?

Searches skills.sh for agent skills you do not have yet, checks their quality, presents the option and installs only after you confirm. When you ask for a skill for some task, or how to do a common task that probably has one, the agent first checks the skills already installed in its context.md if `npx` is missing.

When should I use Agent Skill Finder?

Agent Skill Finder fits situations like: you ask whether a skill exists for a task, or want to add a capability; evaluating a third-party skill's source before installing it; installing a skill from skills.sh into the current project.

How do I install Agent Skill Finder in Claude Code?

Run `npx skills add google/adk-recipes --skill find-skills -a claude-code`. Or copy the skill folder (core/python/long-horizon-harness/horizon/builtin_skills/find-skills in google/adk-recipes) into .claude/skills/find-skills in your project. Claude Code loads it when a task matches its description.

How do I install Agent Skill Finder in Codex?

Run `npx skills add google/adk-recipes --skill find-skills -a codex`. Or copy the skill folder (core/python/long-horizon-harness/horizon/builtin_skills/find-skills in google/adk-recipes) into .agents/skills/find-skills in your project. Codex loads it when a task matches its description.

Can I use Agent Skill Finder in Cursor, Gemini CLI or GitHub Copilot?

Cursor, Gemini CLI, GitHub Copilot and OpenCode also load SKILL.md folders. With the skills CLI, run `npx skills add google/adk-recipes --skill find-skills -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/find-skills, .gemini/skills/find-skills, .github/skills/find-skills and .opencode/skills/find-skills in your project.

What does Agent Skill Finder need to run?

Going by SKILL.md and its folder, Agent Skill Finder needs the command-line tools its instructions call (npx). Our summary lists: Network access to skills.sh; Node.js with npx for the CLI search and install route.

Does Agent Skill Finder access the network?

SKILL.md names 1 domain. As links in the text: skills.sh. This is read from the text; nothing was executed.

Is Agent Skill Finder safe to install?

Our automated static check of SKILL.md found no risky patterns, such as piping downloads into a shell, reading credential files or hidden Unicode. It is not a guarantee. Review the folder before installing.

What licence does Agent Skill Finder use?

Agent Skill Finder is published under the Apache-2.0 licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does Agent Skill Finder use?

About 1.1k tokens (SKILL.md is roughly 4.3k characters). Agents keep only the skill's name and description in context until a task matches; then they load SKILL.md in full.

What are the alternatives to Agent Skill Finder?

Skills that share tags, products or a category with Agent Skill Finder: Skill Base CLI (ginuim/skill-base, 120 stars), TeamAI Team Sync (Tencent/teamai-cli, 5.1k stars), Darwin Skill Optimizer (alchaincyf/darwin-skill, 6.2k stars) and Using Agent Skills (addyosmani/agent-skills, 103k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Agent Skill Finder?

google (a GitHub organization, an official publisher) maintains it in google/adk-recipes, which has 10,421 GitHub stars. The repository holds 14 skills in this directory. The repository was last updated on October 8, 2026.

Source: google/adk-recipes on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.