Agent outcome loop
Let rankings learn from real agent runs.
Resolve
The agent gets one recommended skill.
Resolve returns the selected skill, alternatives, install plan, Trust Score, safety policy, and a unique feedback event id.
Run
The agent tries one narrow task.
Use a sandbox workflow first. Record whether install was used, whether setup was required, and whether risk blocked execution.
Learn
The registry updates trust signals.
Aggregate outcomes improve rankings without exposing raw agent notes or per-user identifiers publicly.
Outcome leaderboard
Skills with the strongest adoption evidence
53 skills currently have reported agent outcomes.
Frontend Design
Early agent signalGuidance for distinctive, intentional UI design, typography, visual direction, and non-template-like product interfaces.
Early agent signal: 2 outcomes, 100% success, 2 install attempts, 0 risk blocks, 0 setup-required reports, Agent Proven Score 51/100.
女娲.Skill
Early agent signal你想蒸馏的下一个员工，何必是同事。蒸馏任何人的思维方式。
Early agent signal: 2 outcomes, 100% success, 2 install attempts, 0 risk blocks, 0 setup-required reports, Agent Proven Score 50/100.
Animation Vocabulary
Early agent signalReverse-lookup glossary that turns a vague description of a web animation or motion effect into its exact term ("the bouncy thing when a popover opens" → Pop in; "the iOS rubber-band scroll" → Rubber-banding). Use when the user asks "what's it called when…", or describes a motion effect without knowing its name and wants the right word to prompt an AI or designer with. For naming an effect, not designing or building one.
Early agent signal: 1 outcome, 100% success, 1 install attempt, 0 risk blocks, 0 setup-required reports, Agent Proven Score 51/100.
Obsidian Second Brain
Early agent signalCross-CLI skill for Obsidian: turn your vault into a living AI-first second brain across Claude Code, Codex, Gemini, and OpenCode. 43 commands - now with /obsidian-architect to document your codebase, key-less web research, Google Calendar, and self-rewriting notes.
Early agent signal: 2 outcomes, 100% success, 1 install attempt, 0 risk blocks, 0 setup-required reports, Agent Proven Score 48/100.
Scrapegraph AI
Early agent signalPython scraper based on AI
Early agent signal: 1 outcome, 100% success, 1 install attempt, 0 risk blocks, 0 setup-required reports, Agent Proven Score 47/100.
Python
Early agent signalAll Algorithms implemented in Python
Early agent signal: 1 outcome, 100% success, 1 install attempt, 0 risk blocks, 0 setup-required reports, Agent Proven Score 42/100.
Markitdown
Early agent signalPython tool for converting files and office documents to Markdown.
Early agent signal: 1 outcome, 100% success, 1 install attempt, 0 risk blocks, 0 setup-required reports, Agent Proven Score 42/100.
MediaCrawler
Early agent signal小红书笔记 | 评论爬虫、抖音视频 | 评论爬虫、快手视频 | 评论爬虫、B站视频 | 评论爬虫、微博帖子 | 评论爬虫、百度贴吧帖子 | 知乎问答文章 | 评论爬虫。支持多平台社交媒体内容抓取，提供完整的数据采集解决方案。
Early agent signal: 1 outcome, 100% success, 1 install attempt, 0 risk blocks, 0 setup-required reports, Agent Proven Score 42/100.
Testcafe
Early agent signalA Node.js tool to automate end-to-end web testing.
Early agent signal: 1 outcome, 100% success, 1 install attempt, 0 risk blocks, 0 setup-required reports, Agent Proven Score 47/100.
Kubernetes Learning Path
Early agent signalA roadmap to learn Kubernetes from scratch (Beginner to Advanced level)
Early agent signal: 1 outcome, 100% success, 1 install attempt, 0 risk blocks, 0 setup-required reports, Agent Proven Score 47/100.
Data Science For Beginners
Early agent signal10 Weeks, 20 Lessons, Data Science for All!
Early agent signal: 1 outcome, 100% success, 1 install attempt, 0 risk blocks, 0 setup-required reports, Agent Proven Score 42/100.
UI-TARS Desktop
Early agent signalRun multimodal agents that operate desktop interfaces
Early agent signal: 1 outcome, 100% success, 1 install attempt, 0 risk blocks, 0 setup-required reports, Agent Proven Score 42/100.
POST /api/agent/outcome
Report after one narrow run.
Resolve responses include a unique feedback.event_id. Agents should reuse it when reporting the result so retries stay idempotent.
{
"event_id": "resolve_...",
"skill_slug": "crawl4ai",
"task": "scrape pricing pages",
"agent": "codex",
"outcome": "success",
"install_used": true,
"time_to_useful_ms": 120000
}Outcome meanings
Use the smallest honest label.
successThe skill helped complete the task.failedThe skill was attempted but did not work.not_relevantThe selected skill did not fit the task.blocked_by_riskAudit, license, token, shell, or network risk stopped execution.setup_requiredThe skill looked relevant but needed missing keys, data, or configuration.