Agent skill

Reference Searcher

by ArabelaTso in ArabelaTso/Skills-4-SE

Systematic external reference searching across documentation, open-source repositories, and web resources.

Apache-2.0Auto-check passedProductivity & Automation

Install Reference Searcher

skills CLI
$ npx skills add ArabelaTso/Skills-4-SE --skill reference-searcher -a claude-code

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

GitHub CLI
$ gh skill install ArabelaTso/Skills-4-SE reference-searcher --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/ArabelaTso/Skills-4-SE.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/reference-searcher .claude/skills/reference-searcher && 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
reference-searcher
GitHub stars
253
Token cost
~1k tokens
SKILL.md length
364 words
Files
1
Skills in repo
170
Repo updated
First seen
Licence
Apache-2.0

At a glance

Systematic external reference searching across documentation, open-source repositories, and web resources.

  • Works in 4 steps: Start with official docs — fastest, most… → If docs insufficient — search GitHub for… → If pattern unclear — web search for… → …
  • Working with unfamiliar libraries
  • SKILL.md covers When to Use This Skill, What This Skill Does, Anti-Patterns and Example
  • Reaches grep.app

What it does

Reference Searcher is an agent skill from ArabelaTso/Skills-4-SE. Systematic external reference searching across documentation, open-source repositories, and web resources. Use when working with unfamiliar libraries, APIs, or when needing production-quality implementation examples.

Its SKILL.md is about 1k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It sits in Productivity & Automation, covering Web search. It works with GitHub. The repository describes itself as: A curated list of 180+ useful Claude Skills for Software Engineering and resources for customizing AI for SE workflows. The licence is Apache-2.0.

When your agent uses it

  • Working with unfamiliar libraries
  • Needing production-quality implementation examples

Example prompts

  • “/reference-searcher”

Workflow steps

4 steps, taken from the first numbered list in SKILL.md.

  1. Start with official docs — fastest, most authoritative
  2. If docs insufficient — search GitHub for real implementations
  3. If pattern unclear — web search for community knowledge
  4. Synthesize — combine findings into actionable guidance

What it can do on your machine

Read from SKILL.md and the folder at commit 4f38503. 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

    No scripts in the folder and no shell commands in SKILL.md (its code samples are bash and markdown).

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

  • Network

    Hosts in commands or code, which the agent is likely to contact:

    • grep.app

    Also links to:

    • github.com

    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

Reference Searcher loads about 1k tokens when it runs. Until then it costs about 59 tokens; SKILL.md has 364 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~59
When it runs · the whole SKILL.md, loaded when a task matches
~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 ArabelaTso/Skills-4-SE at commit 4f38503, republished under its Apache-2.0 licence (© ArabelaTso). 364 words, ~1,028 tokens.

Download SKILL.mdSave it as .claude/skills/reference-searcher/SKILL.md (or your agent's skills folder).
name
reference-searcher
description
Systematic external reference searching across documentation, open-source repositories, and web resources. Use when working with unfamiliar libraries, APIs, or when needing production-quality implementation examples.

Reference Searcher

Structured methodology for searching external references — official docs, open-source implementations, and web resources — to inform development decisions.

When to Use This Skill

  • Working with unfamiliar libraries or frameworks
  • Needing production-quality implementation examples
  • Debugging unexpected behavior from external dependencies
  • Finding best practices for specific patterns (auth, caching, etc.)
  • Evaluating library choices with real-world usage data

What This Skill Does

Search Strategy: Three Layers
Layer 1: Official Documentation

Search official docs for the specific library using web search:

bash
# Search official docs directly
# e.g., "next.js middleware authentication site:nextjs.org"
# Or use documentation MCP tools if available (e.g., Context7)

Best for: API signatures, configuration options, migration guides, official patterns.

Layer 2: Open-Source Implementation Examples (GitHub)

Search real codebases for production patterns:

bash
# Search GitHub for code patterns via grep.app or GitHub search
# e.g., https://grep.app/search?q=getServerSession%28&filter[lang][0]=TypeScript
# Or use GitHub code search: gh search code "getServerSession(" --language=TypeScript

Best for: How established projects (1000+ stars) handle specific patterns, battle-tested error handling, real-world integration examples.

Search tips:

  • Search for actual code patterns, not keywords
  • Use (?s) prefix for multi-line regex patterns
  • Filter by language and file path for precision
  • Look for repos with 1000+ stars for production quality
Layer 3: Web Search (Current Information)

Search the web for recent articles, discussions, and guides:

bash
# Web search for current information
# e.g., "OWASP JWT security best practices 2024"

Best for: Security advisories, recent breaking changes, community discussions, performance benchmarks, "why does X behave this way" questions.

Search Workflow
  1. Start with official docs — fastest, most authoritative
  2. If docs insufficient — search GitHub for real implementations
  3. If pattern unclear — web search for community knowledge
  4. Synthesize — combine findings into actionable guidance
Show full SKILL.md (147 more words)Show less
Result Evaluation

For each source found, assess:

CriterionCheck
RecencyIs this from the current major version?
AuthorityOfficial docs > popular repos > blog posts
CompletenessDoes it cover error handling and edge cases?
ApplicabilityDoes it match our tech stack and constraints?
Output Format
markdown
## Research: [Topic]

### Official Documentation
- [Finding 1 with link]
- [Finding 2 with link]

### Production Examples
- [Repo/file]: [What pattern they use and why]
- [Repo/file]: [Alternative approach]

### Recommendation
Based on [sources], the recommended approach is [X] because [reasons].

### Caveats
- [Known limitation or gotcha]

Anti-Patterns

  • Searching for keywords instead of actual code patterns on GitHub
  • Trusting blog posts over official documentation
  • Using examples from outdated library versions
  • Skipping error handling patterns in reference code
  • Over-researching — stop when you have enough to proceed confidently

Example

User: "How should I implement rate limiting in Express?"

Output:

  1. Official docs: express-rate-limit documentation — configuration options, store adapters
  2. GitHub: 3 production repos showing Redis-backed rate limiting with sliding window
  3. Web: OWASP rate limiting guidelines, comparison of token bucket vs sliding window
  4. Recommendation: Use express-rate-limit with Redis store, sliding window algorithm, with specific config example

Inspired by: oh-my-opencode Librarian agent

© ArabelaTso, 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 skills/reference-searcher of ArabelaTso/Skills-4-SE.

Open the folder on GitHubat commit 4f38503

Compare with similar skills

Reference Searcher 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.

Reference Searcher compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
Reference Searcher this skillArabelaTso/Skills-4-SE253—~1kAutomated safety check: PassApache-2.0
Pi Web Searchsickn33/agentic-awesome-skills47k1 repos~925Automated safety check: PassMIT
Web SearchPhilip-Cao-9527/code-note-helper273—~663Automated safety check: PassMIT
Agent ReachEdisonChenAI/agent-reach1151 repos~1.3kAutomated safety check: PassMIT
Mysearchskernelx/MySearch-Proxy159—~1.4kAutomated safety check: NotesMIT
Z.AI CLInumman-ali/zai-cli110—~528Automated safety check: PassMIT

Similar skills

  • Pi Web Search

    sickn33/agentic-awesome-skills

    Give Pi Agents a safe web-search and fetch workflow using the installed pi-web-access package.

    47k GitHub starsUsed in 1 repo~925 tokens
    Productivity & AutomationAuto-check passed
  • Web Search

    Philip-Cao-9527/code-note-helper

    当任务需要联网检索、外部依据、最新技术知识、官方文档、GitHub 仓库、论文博客、标准、数据集、benchmark、许可证、价格、API 或工程实践依据时使用;适合代码生成前的技术选型、方案设计、内容生成、论文理解、工具库更新和其他 skill 需要外部依据支撑结论的场景。用户明确要求不联网时不要使用。

    273 GitHub stars~663 tokensUpdated 2 mo ago
    Productivity & AutomationAuto-check passed
  • Agent Reach

    EdisonChenAI/agent-reach

    Use the internet: search, read, and interact with 13+ platforms including Twitter/X, Reddit, YouTube, GitHub, Bilibili, XiaoHongShu (小红书), Douyin (抖音), WeChat Articles (微信公众号), LinkedIn, Boss直聘…

    115 GitHub starsUsed in 1 repo~1.3k tokens
    Productivity & AutomationAuto-check passed
  • Mysearch

    skernelx/MySearch-Proxy

    DEFAULT search skill for OpenClaw. An agent skill from skernelx/MySearch-Proxy.

    159 GitHub stars~1.4k tokensUpdated 6 mo ago
    Productivity & AutomationAuto-check: notes
  • Z.AI CLI

    numman-ali/zai-cli

    Command-line access to Z.AI vision analysis, web search, page reading and GitHub repo exploration through npx zai-cli, using an API key.

    110 GitHub stars~528 tokensUpdated 9 mo ago
    Productivity & AutomationAuto-check passed
  • Sponsor Finder

    github/awesome-copilot

    Official

    Find which of a GitHub repository's dependencies are sponsorable via GitHub Sponsors.

    40k GitHub starsUsed in 1 repo~3k tokens
    Productivity & AutomationAuto-check passed

More from ArabelaTso/Skills-4-SE

All 170 skills in this repo
  • Framework Migration Assistant

    ArabelaTso/Skills-4-SE

    Automatically migrate Python web applications between frameworks (Flask → FastAPI, Django → FastAPI).

    253 GitHub stars~1.9k tokensUpdated 1 mo ago
    Auto-check passed
  • Metamorphic Test Generator

    ArabelaTso/Skills-4-SE

    Generate test cases using metamorphic testing by applying transformations based on metamorphic properties.

    253 GitHub stars~798 tokensUpdated 1 mo ago
    Auto-check passed
  • Reproduction Trace Instrumenter

    ArabelaTso/Skills-4-SE

    Instruments programs to capture execution traces specifically for reproducing reported bugs, enabling consistent replay and diagnosis of failures.

    253 GitHub stars~2.4k tokensUpdated 1 mo ago
    Auto-check passed
  • Spring Mvc To Boot Migrator

    ArabelaTso/Skills-4-SE

    Automatically migrate Spring MVC applications to Spring Boot.

    253 GitHub stars~2.2k tokensUpdated 1 mo ago
    Auto-check passed
  • State Snapshot Instrumenter

    ArabelaTso/Skills-4-SE

    Instrument programs (Python, C/C++, Java) to capture snapshots of key program states at runtime, including variables, memory, and call stacks.

    253 GitHub stars~2.2k tokensUpdated 1 mo ago
    Auto-check passed

Works with

Questions about Reference Searcher

What does Reference Searcher do?

Systematic external reference searching across documentation, open-source repositories, and web resources. Reference Searcher is an agent skill from ArabelaTso/Skills-4-SE. Systematic external reference searching across documentation, open-source repositories, and web resources.

When should I use Reference Searcher?

Reference Searcher fits situations like: working with unfamiliar libraries; needing production-quality implementation examples.

How do I install Reference Searcher in Claude Code?

Run `npx skills add ArabelaTso/Skills-4-SE --skill reference-searcher -a claude-code`. Or copy the skill folder (skills/reference-searcher in ArabelaTso/Skills-4-SE) into .claude/skills/reference-searcher in your project. Claude Code loads it when a task matches its description.

How do I install Reference Searcher in Codex?

Run `npx skills add ArabelaTso/Skills-4-SE --skill reference-searcher -a codex`. Or copy the skill folder (skills/reference-searcher in ArabelaTso/Skills-4-SE) into .agents/skills/reference-searcher in your project. Codex loads it when a task matches its description.

Can I use Reference Searcher 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 ArabelaTso/Skills-4-SE --skill reference-searcher -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/reference-searcher, .gemini/skills/reference-searcher, .github/skills/reference-searcher and .opencode/skills/reference-searcher in your project.

What does Reference Searcher need to run?

SKILL.md names no scripts, command-line tools or credentials: Reference Searcher is instructions for the agent only.

Does Reference Searcher access the network?

SKILL.md names 2 domains. In commands or code: grep.app; the agent is likely to contact it when it follows the instructions. As links in the text: github.com. This is read from the text; nothing was executed.

Is Reference Searcher 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 Reference Searcher use?

Reference Searcher 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 Reference Searcher use?

About 1k tokens (SKILL.md is roughly 4.1k 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 Reference Searcher?

Skills that share tags, products or a category with Reference Searcher: Pi Web Search (sickn33/agentic-awesome-skills, 47k stars), Web Search (Philip-Cao-9527/code-note-helper, 273 stars), Agent Reach (EdisonChenAI/agent-reach, 115 stars) and Mysearch (skernelx/MySearch-Proxy, 159 stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains Reference Searcher?

ArabelaTso (a GitHub user) maintains it in ArabelaTso/Skills-4-SE, which has 253 GitHub stars. The repository holds 170 skills in this directory. The repository was last updated on August 21, 2026.

Source: ArabelaTso/Skills-4-SE on GitHub. Facts on this page come from the repository at the commit we read; the author's words are quoted as theirs.