Utilities github-solution-finder
Search GitHub for battle-tested open-source libraries and solutions
install
source · Clone the upstream repo
git clone https://github.com/ouzayr/Utilities
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/ouzayr/Utilities "$T" && mkdir -p ~/.claude/skills && cp -r "$T/DailyPlanner/.local/secondary_skills/github-solution-finder" ~/.claude/skills/ouzayr-utilities-github-solution-finder && rm -rf "$T"
manifest:
DailyPlanner/.local/secondary_skills/github-solution-finder/SKILL.mdsafety · automated scan (low risk)
This is a pattern-based risk scan, not a security review. Our crawler flagged:
- pip install
Always read a skill's source content before installing. Patterns alone don't mean the skill is malicious — but they warrant attention.
source content
GitHub Solution Finder
Find battle-tested libraries instead of building from scratch. Use GitHub's search operators — they're far more precise than plain Google.
Search Operators (combine with spaces = AND)
| Operator | Example | Effect |
|---|---|---|
| | More than N stars |
| | Between N and M |
| | Primary language |
| | Commits after date — the key freshness signal |
| | Repo created after date |
| | Tagged with topic |
| | Specific license |
| | Exclude (prefix any qualifier) |
| Exclude archived repos | |
| No forks | |
/ | | Restrict where term matches |
/ | | Scope to owner |
| | Phrase match |
| | Exclude keyword (strings only) |
High-Signal Query Templates
# Baseline: established + actively maintained <problem> language:<lang> stars:>500 pushed:>2025-06-01 archived:false # Find the dominant library (only a few results = clear winner) <problem> language:python stars:>5000 # Hidden gems (newer, not yet famous, but active) <problem> language:go stars:50..500 pushed:>2025-09-01 fork:false # Curated lists — these exist for almost every topic awesome <topic> in:name stars:>1000 # CLI tools <task> topic:cli stars:>200 pushed:>2025-01-01 # Commercial-safe only <problem> license:mit OR license:apache-2.0 stars:>500 # Boolean grouping (language:rust OR language:go) <problem> stars:>1000 # Code search (different syntax — searches file contents) path:**/*.py "from fastapi import" symbol:RateLimiter
Execute via webSearch
webSearch("site:github.com rate limiter python stars 1000") webSearch("awesome self-hosted analytics github") webSearch("<specific error message> site:github.com/issues")
Note: GitHub-specific qualifiers like
language:, stars:>, and pushed:> only work on GitHub's own search engine. Through webSearch, use natural-language equivalents (e.g. "python" instead of language:python). For precise filtering, use gh search repos if the GitHub CLI is available (see below).
Then
webFetch the repo URL to read the README directly.
Registry alternatives:
site:pypi.org <problem> (Python), site:npmjs.com <problem> (Node), site:crates.io <problem> (Rust).
GitHub CLI (if available)
gh search repos "rate limiter" --language=python --stars=">1000" \ --sort=stars --limit=10 --json=name,stargazersCount,pushedAt,url,description gh api repos/OWNER/REPO --jq '{stars:.stargazers_count, pushed:.pushed_at, issues:.open_issues_count, license:.license.spdx_id, archived:.archived}'
Health Evaluation — Check These Fast
| Signal | Healthy | Walk away |
|---|---|---|
| Last commit | <3 months | >18 months |
| Stars | >1000 (lib), >100 (niche) | <20 |
| Open/closed issue ratio | <0.3 | >1.0 with no replies |
| Contributors | 5+ | 1 (bus factor) |
| "Used by" (sidebar) | >1000 | 0 |
| Releases | Tagged, semver, changelog | No tags |
| License | MIT, Apache-2.0, BSD | None, GPL/AGPL (if commercial) |
| CI badge | Green | Missing or red |
banner | — | Instant no |
Red flags in issues: Search the issue tracker for
"memory leak", "abandoned", "unmaintained", "alternative". If maintainer hasn't replied to anything in 6 months, the project is effectively dead regardless of star count.
Download trend check:
- Python:
— declining = dyinghttps://pypistats.org/packages/<name> - npm:
— compare candidates head-to-headhttps://npmtrends.com/<pkg1>-vs-<pkg2> - Check bundle size:
(frontend only)https://bundlephobia.com/package/<name>
License TL;DR
| License | Commercial OK | Must open-source your code? |
|---|---|---|
| MIT, BSD, Apache-2.0, ISC | Yes | No |
| LGPL | Yes | Only if you modify the lib itself |
| GPL | Yes | Yes, if you distribute (viral) |
| AGPL | Yes | Yes, even for SaaS (network-viral) |
| No LICENSE file | No | default is all rights reserved |
Awesome Lists (curated entry points)
sindresorhus/awesome — the root of all awesome lists. Then: awesome-python, awesome-go, awesome-rust, awesome-react, awesome-selfhosted, awesome-nodejs, free-for-dev, build-your-own-x (learn by reimplementing), public-apis.
Comparison Output Template
## pkg-name [12.4k stars, pushed 2 weeks ago, MIT] github.com/owner/pkg-name **Does:** One-line pitch. **Fit:** Why it matches this specific problem. **Install:** `pip install pkg-name` **Pro:** Active, typed, 89% test coverage. **Con:** Pulls in 23 transitive deps; async-only API. ```python from pkg import Thing Thing().do(x) # minimal working example ```
Decision Rules
- Two libs within 2x stars of each other → pick the one pushed more recently
- A lib with 50k stars but last commit 2023 → it's dead, find the fork (check "Forks" tab sorted by stars)
- Lib does 10x more than needed → check if you can vendor the 200 lines you actually need (with attribution)
- Can't find anything with >100 stars → problem may be too niche; search blog posts / Stack Overflow for how others solved it
- Found 3+ viable options → npmtrends/pypistats comparison, then read the top 5 closed issues of each
Output
Always present key findings and recommendations as a plaintext summary in chat, even when also generating files. The user should be able to understand the results without opening any files.