Agent skill

GitHub Search

by NoobyGains in NoobyGains/godmode

A skill your agent uses when about to build any feature, library, or system - searches GitHub and package registries for existing implementations to study, harvest patterns from, or use directly…

MITAuto-check passed

Install GitHub Search

skills CLI
$ npx skills add NoobyGains/godmode --skill github-search -a claude-code

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

GitHub CLI
$ gh skill install NoobyGains/godmode github-search --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/NoobyGains/godmode.git skills-src && mkdir -p .claude/skills && cp -r skills-src/skills/github-search .claude/skills/github-search && 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
github-search
GitHub stars
107
Token cost
~3.6k tokens
SKILL.md length
1,045 words
Files
1
Skills in repo
34
Repo updated
First seen
Licence
MIT

At a glance

A skill your agent uses when about to build any feature, library, or system - searches GitHub and package registries for existing implementations to study, harvest patterns from, or use directly…

  • Works in 3 steps: Identify Extractable Value → Adapt to Your Context → Attribute the Source
  • About to build any feature
  • SKILL.md covers Overview, The Prime Directive, When to Use and The Entry Protocol, plus 9 more sections
  • Instructions only: no scripts, shell commands, URLs or credentials in SKILL.md

What it does

GitHub Search is an agent skill from NoobyGains/godmode. Use when about to build any feature, library, or system - searches GitHub and package registries for existing implementations to study, harvest patterns from, or use directly instead of building from scratch

Its SKILL.md is about 3.6k tokens, which your agent loads only when the skill is triggered. It is a single SKILL.md file with no bundled scripts.

It works with GitHub. The repository describes itself as: The AI development framework that thinks before it builds. 36 composable skills for Claude Code, Cursor, Codex, and OpenCode. The licence is MIT.

When your agent uses it

  • About to build any feature
  • System - searches GitHub and package registries for existing implementations to study
  • Harvest patterns from
  • Use directly instead of building from scratch

Example prompts

  • “/github-search”

Workflow steps

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

  1. Identify Extractable Value
  2. Adapt to Your Context
  3. Attribute the Source

What it can do on your machine

Read from SKILL.md and the folder at commit 441103a. 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 dot).

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

  • Network

    No URLs in SKILL.md.

    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

GitHub Search loads about 3.6k tokens when it runs. Until then it costs about 55 tokens; SKILL.md has 1,045 words of instructions outside code blocks.

Always · name and description, kept in context so the agent knows when to use it
~55
When it runs · the whole SKILL.md, loaded when a task matches
~3.6k

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 NoobyGains/godmode at commit 441103a, republished under its MIT licence (© NoobyGains). 1,045 words, ~3,624 tokens.

Download SKILL.mdSave it as .claude/skills/github-search/SKILL.md (or your agent's skills folder).
name
github-search
description
Use when about to build any feature, library, or system - searches GitHub and package registries for existing implementations to study, harvest patterns from, or use directly instead of building from scratch

Overview

Before writing a single line of code, determine whether someone on the planet has already solved the same problem and published the result. The strongest engineering move is recognizing when the work has already been done.

Core principle: Search GitHub repositories, package registries, and open-source ecosystems for existing implementations before building anything new. Harvest the best patterns from multiple projects. Attribute what you borrow. Respect licenses.

No exceptions. No workarounds. No shortcuts.

The Prime Directive

NO BUILDING FROM SCRATCH WITHOUT SEARCHING GITHUB FIRST

If you have not queried public repositories for prior solutions, you are burning time on reinvention. Searched three different ways and found nothing applicable? Document what you searched and why nothing fit. Then build.

No excuses:

  • Do not "just start coding and research later"
  • Do not assume uniqueness because the problem feels novel
  • Do not skip research for "straightforward" features (straightforward features have the densest prior art)
  • Do not reject repositories because they are imperfect (extract what is valuable)
  • "I know how to implement this" is not the same as "I should implement this from zero"

When to Use

Mandatory when:

  • Initiating a new feature or project
  • Building something that plausibly exists as open source
  • Weighing competing implementation strategies
  • Seeking architectural precedents from real-world systems
  • Evaluating frameworks, libraries, or tooling choices

Particularly valuable when:

  • The task involves ubiquitous patterns (auth, CRUD, file handling, search, payments)
  • Building a game (every genre has extensive open-source coverage)
  • Creating a utility that addresses a known problem space
  • Implementing algorithms or specialized data structures
  • Setting up infrastructure (CI/CD, deployment pipelines, monitoring)
  • Designing APIs, schemas, or protocols with established conventions

The Entry Protocol

dot
digraph github_search_gate {
    rankdir=TB;
    start [label="Task: Build something", shape=doublecircle];
    search [label="SEARCH\nGitHub, registries,\nopen-source ecosystems", shape=box];
    found [label="Relevant repos found?", shape=diamond];
    evaluate [label="EVALUATE\nStars, maintenance,\ncode quality, license, fit", shape=box];
    usable [label="Directly usable\nas a dependency?", shape=diamond];
    install [label="INSTALL\nAdd as dependency", shape=box style=filled fillcolor=lightgreen];
    harvestable [label="Harvestable\npatterns?", shape=diamond];
    harvest [label="HARVEST\nExtract patterns, adapt\nto your codebase", shape=box style=filled fillcolor=lightyellow];
    studyable [label="Useful as a\nlearning reference?", shape=diamond];
    study [label="STUDY\nAbsorb architecture,\nthen build your own", shape=box];
    build [label="BUILD FROM SCRATCH\nRecord: searched, nothing fit", shape=box];
    record [label="Document search queries\nand findings", shape=box];

    start -> search;
    search -> found;
    found -> evaluate [label="yes"];
    found -> record [label="no"];
    record -> build;
    evaluate -> usable;
    usable -> install [label="yes - strong fit"];
    usable -> harvestable [label="no - too heavy\nor wrong scope"];
    harvestable -> harvest [label="yes - valuable patterns"];
    harvestable -> studyable [label="no - different paradigm"];
    studyable -> study [label="yes - learn from it"];
    studyable -> build [label="no - genuinely nothing relevant"];
}

BEFORE building anything:

  1. SEARCH -- Query GitHub, package registries, and the broader open-source ecosystem
  2. EVALUATE -- Assess candidates for quality, relevance, and license compatibility
  3. DECIDE -- Use as dependency, harvest patterns, study as reference, or build from scratch
  4. DOCUMENT -- Record what you found and why you chose your path
  5. ONLY THEN -- Begin building

Search Methodology

Cast a wide net across multiple channels before concluding nothing exists.

1. GitHub Search: github.com/search (code, repositories, topics)
2. GitHub Topics: github.com/topics/{topic}
3. WebSearch: "github {what you need} {tech stack}"
4. Package registries: npmjs.com, pypi.org, crates.io, rubygems.org, pkg.go.dev
5. Curated lists: awesome-{topic} repositories on GitHub
6. Framework ecosystems: official plugin/extension directories
Effective Querying

Multi-query discipline -- Never settle for a single search.

Query StyleExampleDiscovers
Literal nametetris javascriptExact matches for the concept
Problem statementreal-time collaboration websocketSolutions targeting the same problem
Stack + patternvue authentication oauth2Technology-specific implementations
Synonym explorationkanban board / task tracker / project managerSame concept, different terminology
Curated collectionsawesome-react / awesome-pythonCommunity-vetted lists

Minimum threshold before declaring "nothing exists": 3 distinct query formulations across at least 2 search channels.

What to Search For
dot
digraph search_scope {
    "What am I building?" [shape=diamond];
    "Complete application?" [shape=diamond];
    "Feature or module?" [shape=diamond];
    "Search for full repos" [shape=box];
    "Search for libraries + example repos" [shape=box];
    "Search for code snippets + patterns" [shape=box];

    "What am I building?" -> "Complete application?" [label="assess"];
    "Complete application?" -> "Search for full repos" [label="yes"];
    "Complete application?" -> "Feature or module?" [label="no"];
    "Feature or module?" -> "Search for libraries + example repos" [label="yes"];
    "Feature or module?" -> "Search for code snippets + patterns" [label="no"];
}

Evaluating Repositories

When assessing discovered repositories, apply these signals systematically.

SignalPositive IndicatorNegative Indicator
Popularity100+ stars (community-validated)0-5 stars (unvetted)
MaintenanceCommit within 6 monthsDormant 2+ years
Community healthActive discussions, merged PRsHundreds of stale issues, no maintainer response
LicenseMIT, Apache 2.0, BSDGPL (if you need permissive), no license at all
Code disciplineTests, types, documentation, clean structureNo tests, no types, tangled code
Dependency footprintMinimal, well-known dependencies50+ transitive deps, obscure packages
DocumentationClear setup instructions, usage examplesEmpty or outdated README
RelevanceAddresses 70%+ of your needTangentially related

Minimum viable candidate: Has tests, has a license, updated within the past year, README explains usage.

Quick Evaluation Checklist
For each candidate repository:

1. STARS & FORKS: Gauge community trust (100+ stars is a reasonable floor)
2. LAST COMMIT: Anything older than 18 months is a maintenance risk
3. OPEN ISSUES: High count with no maintainer replies signals abandonment
4. LICENSE FILE: No license means you cannot legally use it
5. TEST DIRECTORY: No tests means no confidence in correctness
6. README QUALITY: Poor docs usually correlate with poor internal structure
7. DEPENDENCIES: Check for bloated or unmaintained transitive deps
8. RELEASE CADENCE: Regular releases indicate ongoing investment

Pattern Harvesting

When a repository is not directly consumable but contains valuable patterns, extract intelligently.

Phase 1: Identify Extractable Value
FROM the repository, extract:
- Architectural patterns (how they organized the project)
- Algorithm implementations (how they solved the hard parts)
- Data models (how they structured domain entities)
- Interface designs (how they shaped the public API)
- Error handling approaches (how they manage edge cases)
- Test strategies (how they verify similar functionality)
Phase 2: Adapt to Your Context
DO NOT copy-paste entire files.
DO extract the pattern and rewrite for your:
- Technology stack
- Naming conventions
- Design system (for UI code)
- Architectural layering
- Internal codebase conventions (use godmode:codebase-research to find them)
Phase 3: Attribute the Source
In your codebase or documentation, note:
"Approach inspired by github.com/author/repo - [what was adapted]"

License Compatibility

Before using any code from a discovered repository, verify license compatibility.

Your Project LicenseCompatible Source Licenses
MITMIT, BSD, Apache 2.0, ISC, Unlicense
Apache 2.0MIT, BSD, Apache 2.0, ISC, Unlicense
GPLAny (GPL is permissive for incoming code)
ProprietaryMIT, BSD, Apache 2.0, ISC, Unlicense
AnyNever use: No license stated, AGPL (unless you comply fully)

When uncertain: MIT and Apache 2.0 are safe for virtually any project. GPL requires your project to also be GPL. No license means all rights reserved by the author -- do not use.

Show full SKILL.md (391 more words)Show less

Multi-Source Harvesting

For complex features, the optimal approach often combines patterns from several repositories.

Example: Building a real-time collaborative editor

Repo 1 (github.com/x/rich-editor): Polished prosemirror integration
  -> Harvest: Editor initialization pattern, schema definition approach

Repo 2 (github.com/y/crdt-sync): Clean CRDT implementation for text
  -> Harvest: Conflict resolution algorithm, operation transformation logic

Repo 3 (github.com/z/ws-rooms): WebSocket room management
  -> Harvest: Connection lifecycle, reconnection strategy, presence tracking

Result: Your implementation synthesizes the strongest elements from 3 repos,
each designed by specialists in their domain.

Report to the user:

Present discoveries as labeled options with links. Mark the recommendation with a star.

I found [N] relevant repositories. Here are the top candidates:

**A)** [repo-name] — [stars] stars, [license]
   Link: [GitHub URL]
   Extractable: [specific patterns/code worth harvesting]
   Fit: [what percentage of your need it covers]

**B)** [repo-name] — [stars] stars, [license]
   Link: [GitHub URL]
   Extractable: [specific patterns/code worth harvesting]
   Fit: [what percentage of your need it covers]

**C)** [repo-name] — [stars] stars, [license]
   Link: [GitHub URL]
   Extractable: [specific patterns/code worth harvesting]
   Fit: [what percentage of your need it covers]

**D) Multi-source harvest** — Best patterns from all three ⭐ Recommended
   From A: [what to extract]
   From B: [what to extract]
   From C: [what to extract]
   Build from scratch: [what is unique to this project]
   Why: Combines battle-tested patterns from [N] production systems

Check out the repos and pick A, B, C, or D.

YoloMode exception: Reference and repository selection is ALWAYS interactive — even in YoloMode, present the options and let the user pick. These choices are too impactful to auto-select.

Cognitive Traps

RationalizationTruth
"I can build it faster than studying someone else's code"You will also maintain it indefinitely. Battle-tested code has fewer defects.
"Nothing exists for my exact scenario"Did you try 3+ query variations? Partial matches are valuable.
"Open source quality is unreliable"Repositories with 1k+ stars and test suites are often superior to what you will produce under time pressure.
"It is faster to just start writing"You will spend hours solving problems someone already addressed.
"I want to avoid external dependencies"Harvest patterns without adding dependencies. The knowledge is free.
"Licensing is too complicated"MIT/Apache/BSD means free to use. A two-second check saves you from reinvention.
"It is only 100 lines, not worth researching"Those 100 lines with edge-case coverage someone already wrote outperform your 100 lines without it.
"I will research if I get stuck"Research FIRST. Getting stuck means you already burned time.

Guardrails

Prohibited actions:

  • Starting implementation without searching first (minimum 3 queries)
  • Rejecting all search results without evaluation
  • Copying code without verifying license compatibility
  • Using a repository with no license (legal risk)
  • Assuming your solution will surpass a battle-tested one
  • Omitting attribution when harvesting patterns

Required actions:

  • Search with at least 3 different query formulations
  • Evaluate candidates against the assessment criteria
  • Verify license compatibility before using any code
  • Document what you found and your rationale
  • Harvest patterns even from repositories you do not use directly
  • Report findings to the user with links, popularity metrics, and what you are extracting

Quick Reference

SEARCH -> EVALUATE -> DECIDE -> DOCUMENT -> BUILD

Search: 3+ queries across GitHub, registries, and curated lists
Evaluate: Stars, activity, tests, license, relevance
Decide: Use as dependency | Harvest patterns | Study as reference | Build from scratch
Document: What you searched, what you found, why you chose your path
Build: With patterns from research, not from assumptions

Integration

Invoked during:

  • godmode:intent-discovery -- Search during "Explore project context" and research phases
  • godmode:reference-engine -- Routed here for external code and library research
  • godmode:task-planning -- Reference repositories in plan tasks
  • godmode:system-design -- Discover reference architectures on public repositories

Complementary skills:

  • godmode:codebase-research -- For searching WITHIN the current codebase for internal patterns, conventions, and existing implementations to match
  • godmode:specification-first -- Feeds search findings into formal specifications
  • godmode:ux-patterns -- Repositories for code patterns; UX system for design patterns
  • godmode:design-integration -- Find design system repos for bootstrapping
  • godmode:project-bootstrap -- Locate starter templates and boilerplate projects

© NoobyGains, MIT. 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/github-search of NoobyGains/godmode.

Open the folder on GitHubat commit 441103a

Compare with similar skills

GitHub Search 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.

GitHub Search compared with similar skills
SkillStarsUsed inTokensAuto-checkLicenceRepo updated
GitHub Search this skillNoobyGains/godmode107—~3.6kAutomated safety check: PassMIT
PR Babysitteropeninterpreter/openinterpreter69k3 repos~4.2kAutomated safety check: PassApache-2.0
Diagnosing Superpowers Sessionsobra/superpowers296k3 repos~1.7kAutomated safety check: PassMIT
GitHub Deep Researchbytedance/deer-flow83k5 repos~1.3kAutomated safety check: PassMIT
Greplooponyx-dot-app/onyx32k4 repos~3.3kAutomated safety check: PassMIT
Update V8 Versionopeninterpreter/openinterpreter69k2 repos~845Automated safety check: PassApache-2.0

Similar skills

  • PR Babysitter

    openinterpreter/openinterpreter

    Watches an open GitHub pull request until it merges, handling review comments, diagnosing CI failures and retrying flaky checks along the way.

    69k GitHub starsUsed in 3 repos~4.2k tokens
    DevelopmentAuto-check passed
  • Investigates a session where Superpowers went wrong, reads the transcripts on disk and produces an evidence-cited report, optionally prepared as a bug report for the maintainers.

    296k GitHub starsUsed in 3 repos~1.7k tokens
    Agent WorkflowsAuto-check passed
  • GitHub Deep Research

    bytedance/deer-flow

    Researches a GitHub repository over four rounds using the GitHub API and web search, then writes a structured markdown report with timeline, metrics and Mermaid diagrams.

    83k GitHub starsUsed in 5 repos~1.3k tokens
    Research & ScienceAuto-check passed
  • Greploop

    onyx-dot-app/onyx

    Iteratively improves a PR (GitHub), MR (GitLab), or shelved changelist (Perforce) until Greptile gives it a 5/5 confidence score with zero unresolved comments.

    32k GitHub starsUsed in 4 repos~3.3k tokens
    DevelopmentAuto-check passed
  • Update V8 Version

    openinterpreter/openinterpreter

    Bumps the pinned v8 and rusty_v8 versions in Codex, validates the release-candidate path with the v8-canary check, and traces failures to upstream build changes.

    69k GitHub starsUsed in 2 repos~845 tokens
    DevOps & CloudAuto-check passed
  • Check PR

    onyx-dot-app/onyx

    Checks a GitHub, GitLab, or Perforce (p4) pull request (or merge request, or shelved changelist) for unresolved review comments, failing status checks, and incomplete PR descriptions.

    32k GitHub starsUsed in 2 repos~2.3k tokens
    DevelopmentAuto-check passed

More from NoobyGains/godmode

All 34 skills in this repo
  • Activation

    NoobyGains/godmode

    A skill your agent uses when starting any conversation - establishes how to locate and invoke skills, mandating Skill tool usage before ANY response including clarifying questions

    107 GitHub stars~2.4k tokensUpdated 7 mo ago
    Auto-check passed
  • Agent Messaging

    NoobyGains/godmode

    A skill your agent uses when dispatching subagents, composing prompts for teammates, structuring handoff reports, or managing context boundaries between agents.

    107 GitHub stars~3k tokensUpdated 7 mo ago
    Auto-check passed
  • Codebase Research

    NoobyGains/godmode

    A skill your agent uses when building ANY feature within an existing project - search the current codebase for existing patterns, conventions, similar implementations, and established approaches…

    107 GitHub stars~3.2k tokensUpdated 7 mo ago
    Auto-check: notes
  • Completion Gate

    NoobyGains/godmode

    A skill your agent uses when about to declare work done, fixed, or passing, before committing or opening PRs - demands executing verification commands and reading their output before making any…

    107 GitHub stars~1.6k tokensUpdated 7 mo ago
    Auto-check passed
  • Comprehension Check

    NoobyGains/godmode

    A skill your agent uses when implementing any substantial feature, multi-file modification, or architectural change - produces a plain-language walkthrough of every alteration so the developer can…

    107 GitHub stars~1.5k tokensUpdated 7 mo ago
    Auto-check passed
  • Delegated Execution

    NoobyGains/godmode

    A skill your agent uses when executing implementation plans with independent tasks in the current session

    107 GitHub stars~2.4k tokensUpdated 7 mo ago
    Auto-check passed

Works with

Questions about GitHub Search

What does GitHub Search do?

A skill your agent uses when about to build any feature, library, or system - searches GitHub and package registries for existing implementations to study, harvest patterns from, or use directly…. GitHub Search is an agent skill from NoobyGains/godmode.

When should I use GitHub Search?

GitHub Search fits situations like: about to build any feature; system - searches GitHub and package registries for existing implementations to study; harvest patterns from; use directly instead of building from scratch.

How do I install GitHub Search in Claude Code?

Run `npx skills add NoobyGains/godmode --skill github-search -a claude-code`. Or copy the skill folder (skills/github-search in NoobyGains/godmode) into .claude/skills/github-search in your project. Claude Code loads it when a task matches its description.

How do I install GitHub Search in Codex?

Run `npx skills add NoobyGains/godmode --skill github-search -a codex`. Or copy the skill folder (skills/github-search in NoobyGains/godmode) into .agents/skills/github-search in your project. Codex loads it when a task matches its description.

Can I use GitHub Search 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 NoobyGains/godmode --skill github-search -a cursor` (or -a gemini-cli, github-copilot or opencode for the others). To copy it by hand, put the folder in .cursor/skills/github-search, .gemini/skills/github-search, .github/skills/github-search and .opencode/skills/github-search in your project.

What does GitHub Search need to run?

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

Does GitHub Search access the network?

SKILL.md contains no URLs. Any network use would come from the scripts or tools the agent runs. This is read from the text; nothing was executed.

Is GitHub Search 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 GitHub Search use?

GitHub Search is published under the MIT licence (the repository's licence). It allows redistribution, so the full SKILL.md is shown on this page.

How many tokens does GitHub Search use?

About 3.6k tokens (SKILL.md is roughly 14k 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 GitHub Search?

Skills that share tags, products or a category with GitHub Search: PR Babysitter (openinterpreter/openinterpreter, 69k stars), Diagnosing Superpowers Sessions (obra/superpowers, 296k stars), GitHub Deep Research (bytedance/deer-flow, 83k stars) and Greploop (onyx-dot-app/onyx, 32k stars). The comparison table on this page puts their stars, adoption, token cost, safety result and licence side by side.

Who maintains GitHub Search?

NoobyGains (a GitHub user) maintains it in NoobyGains/godmode, which has 107 GitHub stars. The repository holds 34 skills in this directory. The repository was last updated on March 9, 2026.

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