Agent outcome loop
Resolve
Resolve returns the selected skill, alternatives, install plan, Trust Score, safety policy, and a unique feedback event id.
Run
Use a sandbox workflow first. Record whether install was used, whether setup was required, and whether risk blocked execution.
Learn
Aggregate outcomes improve rankings without exposing raw agent notes or per-user identifiers publicly.
Outcome leaderboard
222 skills currently have reported agent outcomes.
Guidance for distinctive, intentional UI design, typography, visual direction, and non-template-like product interfaces.
Promising agent evidence: 47 outcomes, 94% success, 47 install attempts, 0 risk blocks, 0 setup-required reports, Agent Proven Score 70/100.
Generate original one-ink or controlled two-ink editorial images from any theme, sentence, article idea, object, or reference photo. Always use this skill when the user asks for 单色海报、双色印刷、单色调视觉、蓝色/绿色孔版印刷、risograph、网点照片、复古或当代编辑排版、zine poster, monochrome editorial poster, duotone print, or asks to use the mono-color style. It uses an adaptive white, gray, or pale-beige substrate, no more than two printing inks, active negative space, terse human language, and strong serif/grotesk/mono typography without making retro styling the default or copying a source composition, wording, logo, or artwork. Produce both the final generation prompt and the generated raster image unless the user explicitly asks for prompt only.
Promising agent evidence: 45 outcomes, 100% success, 45 install attempts, 0 risk blocks, 0 setup-required reports, Agent Proven Score 80/100.
Turn one topic into a narrated Vox-style paper-collage explainer or ad video, from script through captions.
Promising agent evidence: 31 outcomes, 100% success, 31 install attempts, 0 risk blocks, 0 setup-required reports, Agent Proven Score 78/100.
Design and implementation guidance for distinctive landing pages, portfolios, product demos, and purposeful redesigns.
Early agent signal: 28 outcomes, 86% success, 28 install attempts, 0 risk blocks, 0 setup-required reports, Agent Proven Score 65/100.
AI-agent Skill for generating polished HTML slide decks: editorial magazine and Swiss layouts, image prompts, social covers, and a WebGL/low-power presentation runtime.
Promising agent evidence: 20 outcomes, 100% success, 20 install attempts, 0 risk blocks, 0 setup-required reports, Agent Proven Score 74/100.
Any agent Skill: generate beautiful architecture diagrams with dark/light theme toggle and PNG/JPEG/WebP/SVG export
Promising agent evidence: 17 outcomes, 100% success, 17 install attempts, 0 risk blocks, 0 setup-required reports, Agent Proven Score 75/100.
An agentic skills framework & software development methodology that works.
Promising agent evidence: 13 outcomes, 92% success, 13 install attempts, 0 risk blocks, 0 setup-required reports, Agent Proven Score 69/100.
Create original visual art, posters, PNG assets, and PDF documents through a clear design philosophy.
Promising agent evidence: 9 outcomes, 100% success, 9 install attempts, 0 risk blocks, 0 setup-required reports, Agent Proven Score 74/100.
Create beautiful slides on the web using a coding agent's frontend skills
Early agent signal: 11 outcomes, 91% success, 10 install attempts, 0 risk blocks, 0 setup-required reports, Agent Proven Score 67/100.
Use Playwright to interact with and test local web applications, capture screenshots, debug UI behavior, and inspect browser logs.
Promising agent evidence: 8 outcomes, 100% success, 8 install attempts, 0 risk blocks, 0 setup-required reports, Agent Proven Score 73/100.
Review UI code for web interface guidelines, UX quality, accessibility, and interaction design best practices.
Promising agent evidence: 8 outcomes, 100% success, 8 install attempts, 0 risk blocks, 0 setup-required reports, Agent Proven Score 73/100.
🎨 Local-first, open-source Claude Design alternative. 🖥️ Native desktop app. ⚡ 259+ Skills · ✨ 142+ Design Systems 🖼️ Web · desktop · mobile prototypes · slides · images · videos · HyperFrames 📦 Sandboxed preview · HTML/PDF/PPTX/MP4 export 🤖 Claude Code / OpenClaw / Codex / Cursor / OpenCode / Qwen / Copilot / Hermes / Kimi & 17+ CLIs.
Early agent signal: 8 outcomes, 88% success, 8 install attempts, 0 risk blocks, 0 setup-required reports, Agent Proven Score 67/100.
POST /api/agent/outcome
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
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.