Skills Library
github.com/dylanfeltus/skills- Category
- AI Agents
- Rank
- No. 1069Tools index
Previous survey · No. 1128 ·
- Pricing
- Open Source
- Platform
- cli
- Type
- TOOL
- Builder
- @dylanfeltus
- GitHub
- 179 stars
- Date
About
A collection of reusable AI agent skills that provide domain expertise and systematic workflows for research, design, and content creation. Instead of prompting from scratch, agents can load pre-built skills for consistent, high-quality output across various tasks like trademark searches, design patterns, and content improvement.
What it does
Skills Library supplies installable instruction folders for AI agents. Each folder defines when a task applies, which external services or tools to use, expected output, and failure handling. Users can install the whole set or select one skill, while agents follow the included Markdown workflow.
Why it's ranked here
The strongest material turns vague requests into explicit procedures, especially for design tokens, visual checks, and API research. Examples, response shapes, fallbacks, and error cases make the guidance practical. Quality still varies because several workflows depend on search results, scraping, model judgment, or external services rather than verified local code.
What's good
The instructions expose important constraints instead of hiding them. Product Hunt documents its lack of text search. Trademark research warns that search snippets are incomplete and not legal advice. Virtual-card guidance limits spending and protects card details. Design guidance includes formulas, thresholds, output formats, and ordered checks that an agent can apply consistently.
Tradeoffs
This is primarily Markdown guidance, not an executable toolkit with shared validation or automated tests shown in the supplied text. Results depend on an agent supporting the named browsing, vision, shell, or image capabilities. Product Hunt and virtual cards require credentials. Google Play relies on fragile scraping, while trademark checks rely mainly on search-engine indexing.
How to use it well
It suits teams already using compatible coding agents and wanting repeatable playbooks for narrow, recurring tasks. Install only the relevant folders, inspect credential and tool requirements, then verify consequential outputs against primary sources. It does not replace legal counsel, payment authorization controls, stable data integrations, or executable test coverage for generated design and research results.
Technical notes+
README.md documents whole-library or selective installation through npx skills add dylanfeltus/skills, plus manual folder copying. Each shown module pairs a SKILL.md instruction contract with a human-facing README.md. hn-search/SKILL.md specifies Algolia query construction and response handling; producthunt/SKILL.md uses authenticated GraphQL with web-search fallback; appstore-intel/SKILL.md combines Apple lookup endpoints with Google Play scraping; visual-qa/SKILL.md expects screenshot and vision tooling; privacy-cards/SKILL.md contains REST workflows and sensitive-data controls; design-tokens/SKILL.md provides formulas and CSS, Tailwind, and JSON output patterns; trademark-search/SKILL.md relies on indexed web results because direct database fetches are described as blocked.
Observed
- License
- MIT is stated in the supplied per-skill README documents.
- Primary format
- Markdown instruction folders with agent-facing and human-facing documentation.
- Install surface
- Install all or selected skills through npx, or copy folders manually.
- Agent compatibility
- README names OpenClaw, Claude Code, Cursor, Codex, and other LLM-based tools.
- Interfaces
- Instruction library for agents; individual skills direct HTTP APIs, web search, scraping, shell commands, vision tools, or browser tools.
- Authentication
- Several skills need no key; Product Hunt and Privacy.com workflows require credentials.
- Repository structure
- The supplied tree shows paired skill instructions and README documentation, with no test directory or package manifest shown.
Read from README.md, AGENTS.md, hn-search/SKILL.md, visual-qa/SKILL.md, hn-search/README.md, visual-qa/README.md, producthunt/SKILL.md, producthunt/README.md, design-tokens/SKILL.md, privacy-cards/SKILL.md, appstore-intel/SKILL.md, design-tokens/README.md, privacy-cards/README.md, appstore-intel/README.md, trademark-search/SKILL.md.
What it can do
Load pre-built skills into AI agents
AI agent system and skill selection → Agent with enhanced domain expertise capabilities
Perform trademark searches
Trademark query parameters → Structured trademark search results
Generate design patterns
Design requirements and constraints → Systematic design pattern recommendations
Improve content quality
Raw content text → Enhanced, refined content
Execute research workflows
Research topic and parameters → Structured research findings and analysis
Provide systematic content creation workflows
Content creation requirements → Step-by-step content creation process
Tags
Comments (0)
No comments yet
Editorially curated, with community endorsements as a secondary signal. Corrections welcome.