OpenAgentSkill (@openagentskill) header
OpenAgentSkill (@openagentskill) profile picture

OpenAgentSkill

@openagentskill

The best agent skill store🦞 Discover, compare, and install reusable skills across Codex, Claude Code, Cursor, and agent workflows.

 Agent world
openagentskill.com
Joined September 2023
  • Tweets541
  • Following16
  • Followers782
find it on openagentskill.com
Image from OpenAgentSkill (@openagentskill)
16
codex made it !
1
2
85
cool! already list on openagentskill.👇 openagentskill.com/skills/de…
Image from OpenAgentSkill (@openagentskill)
Leon Lin (@LexnLin) profile pictureLeon Lin
@LexnLinOct 13
We rebuilt TasteSkill from scratch And today we are introducing TasteSkill v2 v2 gives your coding agent an art direction, assets and a browser check to verify its work npx skills add Leonxlnx/taste-skill check out more examples below!
1
45
New Skills Added
1
2
3
92
openagentskill.com/showcase/…
Image from OpenAgentSkill (@openagentskill)
14
Openagentskill Update GC Minimal Zine Poster: three quiet studies by LiamGvchi
Image from OpenAgentSkill (@openagentskill)
1
1
41
Your agent rules may be in the wrong place. AGENTS.md: project context. Skills: repeatable procedures. MCP: connected capabilities. A practical code-review example, plus checks before adoption: openagentskill.com/blog/agen…
Three columns distinguish AGENTS.md for repository context, Skills for reusable task procedures, and MCP for connected tools and context. Recommendation: keep repository-specific rules with the repository. Documentation review, not a runtime benchmark.

ALT Three columns distinguish AGENTS.md for repository context, Skills for reusable task procedures, and MCP for connected tools and context. Recommendation: keep repository-specific rules with the repository. Documentation review, not a runtime benchmark.

Migration review example: AGENTS.md contains the local rollback-note requirement; a skill defines evidence-based review steps; an authorized MCP integration retrieves current CI results. Instructions do not grant deployment or merge permission. This is a proposed workflow.

ALT Migration review example: AGENTS.md contains the local rollback-note requirement; a skill defines evidence-based review steps; an authorized MCP integration retrieves current CI results. Instructions do not grant deployment or merge permission. This is a proposed workflow.

Proposed review pilot with three checks: cite the actual repository rule, give an investigable finding with evidence, and report unavailable CI as unverified. Verify instruction loading and tool permissions in the actual host. No runtime test was performed.

ALT Proposed review pilot with three checks: cite the actual repository rule, give an investigable finding with evidence, and report unavailable CI as unverified. Verify instruction loading and tool permissions in the actual host. No runtime test was performed.

1
1
75
NEW ~ Skill Gallery
Image from OpenAgentSkill (@openagentskill)
1
37
Now, Send your skills on openagentskill ！！
Image from OpenAgentSkill (@openagentskill)
3
38
We’re making AI agent skills truly global. Your agents can now discover, compare, and install the best reusable skills in their preferred language. Supported languages: English (EN) Chinese (ZH) Japanese (JA) Korean (KO) Spanish (ES) German (DE) French (FR) Indonesian (ID) And more languages coming soon! We’re committed to becoming the world’s highest-quality skill platform for AI agents — with the largest indexed library, seamless auto-installs, and the smartest recommendation system. #OpenAgentSkill #AISkills #MultiLanguage #AIAgents
Image from OpenAgentSkill (@openagentskill)
1
3
141
AI agents need a skill layer. OpenAgentSkill helps agents find, compare, audit, and install reusable skills automatically. npm for AI Agent Skills.
Image from OpenAgentSkill (@openagentskill)
1
1
37
AI agents should not choose skills from a static directory. They need a registry that can answer: What skill should I use? Can I trust it? How do I install it? What happened when other agents tried it? We shipped the Agent Outcome Loop. openagentskill.com/outcomes
🤖 Made with AI
Image from OpenAgentSkill (@openagentskill)
Image from OpenAgentSkill (@openagentskill)
Image from OpenAgentSkill (@openagentskill)
Image from OpenAgentSkill (@openagentskill)
1
1
64
Agents should not install random GitHub repos blindly. Every skill should carry trust signals before execution: quality, maintenance, license, install safety, risk level, and alternatives. That is why OpenAgentSkill is building Trust Score + audit pages for skills.
Image from OpenAgentSkill (@openagentskill)
1
1
32
Today, most agents still start from a blank prompt. But real work needs reusable capabilities: crawl a site, analyze stock news, parse PDFs, review code, build reports. OpenAgentSkill turns a task into a ranked skill plan: best skill, alternatives, install path, and risk notes.
Image from OpenAgentSkill (@openagentskill)
1
31
OpenAgentSkill is the skill layer for AI agents: find, compare, audit, and install reusable skills automatically. 20,000+ indexed skills Trust Score + audit pages Built for Codex, Claude Code, Cursor
Image from OpenAgentSkill (@openagentskill)
2
5
442
OpenAgentSkill helps agents: 1. Discover skills by task 2. Compare quality, stars, maintenance, and trust 3. Get install commands and API-ready metadata 4. Compose workflows faster
Image from OpenAgentSkill (@openagentskill)
1
3
38
Load more