独立コード監査＆ガバナンス評価ラボ
Agent Skill: 監査・統治・ベンチマーク評価
単なるプロンプト集ではなく、静的コード解析を通じてオープンソースAgent Skillの内部構造を解剖し、L1〜L5の実行権限境界を測定して実用的な推奨を発行します。
★ FLAGSHIP COMPARISON · 175,000+ STARS
累計17.5万 GitHub Stars！中国トップAIブロガーによるオープンソースAgent Skill決定版（ビジュアル作例＆ワンクリック導入ガイド付き）
AI特有の「安っぽさ」や「プラスチック感」から今すぐ脱却！宝玉、李継剛、帰蔵、小互、数字生命卡茲克、花叔、向陽喬木、黄叔、蒼何、小黒の10名が公開する強力なオープンソースAgent Skillを徹底解剖。累計17.5万超のGitHub Starsを誇る実績、各クリエイターの設計思想、ビジュアル生成作例、導入コマンドを網
STARTER PACK · ZERO RISK
最近話題のAI画像加工・美化Skill 6選（まとめ転載）
転載：GitHubでオープンソース公開されている注目のAI画像加工Skill 6選——ZINEポスター、ピクセル印刷、引き算の美学、淡墨版画、旅行抽象化、ちぎり絵コラージュ。すべて当サイトに収録済み＆ソースコード検証済み。
VISUAL DISCIPLINE
初心者がまず導入すべき10選のSkills
初めてAgentにSkillを導入する人へ：公式提供や圧倒的な高評価を誇るオープンソースSkill 10選。「思考の整理 → タスク化 → 学習 → UI作成 → リサーチ → ドキュメント作成」の全プロセスを網羅。
guizang-social-card-skill
@op7418/guizang-social-card-skill · Commit cf4b810
Verified Output
Architectural Analysis & Mechanism Verdict
Re-engineers graphic layout into a deterministic recipe: M01–M16 visual models coupled with 9 headless Playwright assertion rules. Enforces zero synthetic data, bounded typography, and strict 21:9 to 1:1 paired cover generation.
Host Security Governance & Runtime Boundaries
Host Privilege: L4 · L4 高権限
Scripts Executed: 5 个辅助子进程
Network Egress: 4 个外呼网络端点
Runtime Policy: 分级授权与审计
baoyu-infographic
@JimLiu/baoyu-skills · Commit 1567581
Verified Output
Architectural Analysis & Mechanism Verdict
Deconstructs infographics into orthogonal dimensions: 21 layout structures × 22 aesthetic styles, offering modular combinations far exceeding static templates. Implements a mandatory confirmation gate preventing unapproved image generation.
Host Security Governance & Runtime Boundaries
Host Privilege: L4 · L4 高権限
Scripts Executed: 1 个辅助子进程
Network Egress: 2 个外呼网络端点
Runtime Policy: 分级授权与审计
khazix-writer
@KKKKhazix/khazix-skills · Commit 48e8ba5
Verified Output
Architectural Analysis & Mechanism Verdict
Engineers human-like writing rhythm: replaces bloated corporate prose with conversational cadences, sudden breaks, and 12 narrative patterns. Features a 4-tier automated audit scanning against punctuation cliches and AI tells.
Host Security Governance & Runtime Boundaries
Host Privilege: L1 · L1 基本
Scripts Executed: 0 个辅助子进程
Network Egress: 0 个外呼网络端点
Runtime Policy: 登记来源即可
design-taste-frontend
@leonxlnx/taste-skill · Commit ccbc156
Verified Output
Architectural Analysis & Mechanism Verdict
Actionable anti-slop frontend mechanics: strictly enforces bans on em-dashes, caps hero viewport heights, limits zigzag sections, and executes a pre-flight checklist against synthetic UI artifacts.
Host Security Governance & Runtime Boundaries
Host Privilege: L1 · L1 基本
Scripts Executed: 0 个辅助子进程
Network Egress: 0 个外呼网络端点
Runtime Policy: 登记来源即可
scenes-gathered-zine
@Zeejay0/gathered-scenes-zine-skill · Commit b9edb83
Verified Output
Architectural Analysis & Mechanism Verdict
Structured visual decision pipeline: Scene Card → Abstraction Map → 5 illustration syntaxes → paper texture treatment. Input/output contracts are strictly bounded, preventing sloppy collage outputs.
Host Security Governance & Runtime Boundaries
Host Privilege: L3 · L3 ネットワーク
Scripts Executed: 0 个辅助子进程
Network Egress: 2 个外呼网络端点
Runtime Policy: 声明域名与密钥范围
youtube-downloader
@daymade/claude-code-skills · Commit 63a65c1
Verified Output
Architectural Analysis & Mechanism Verdict
Highly engineered anti-scraping resilience: encodes known failure modes (PO token, 403, SSL EOF) into automated classification and retry chains for deterministic local media extraction.
Host Security Governance & Runtime Boundaries
Host Privilege: L5 · L5 厳格管理
Scripts Executed: 3 个辅助子进程
Network Egress: 6 个外呼网络端点
Runtime Policy: 隔离环境与严格审计
grill-me
@mattpocock/skills · Commit 3cca18b
Verified Output
Architectural Analysis & Mechanism Verdict
Thin-shell Socratic interrogation: transforms interview primitives into interactive feedback loops to stress-test designs and architectural plans before code implementation.
Host Security Governance & Runtime Boundaries
Host Privilege: L1 · L1 基本
Scripts Executed: 0 个辅助子进程
Network Egress: 0 个外呼网络端点
Runtime Policy: 登记来源即可
baoyu-comic
@JimLiu/baoyu-skills · Commit 1567581
Verified Output
Architectural Analysis & Mechanism Verdict
Expansive narrative composition: 6 art styles × 7 emotional tones × 7 panel layouts with specialized narrative constraints (Ohmsha technical metaphor, 4-panel progression).
Host Security Governance & Runtime Boundaries
Host Privilege: L4 · L4 高権限
Scripts Executed: 1 个辅助子进程
Network Egress: 0 个外呼网络端点
Runtime Policy: 分级授权与审计
L1 Pure Instruction 164 Skills
Prompt-only specifications with zero scripts, zero network egress, and zero file access. Completely safe for any host runtime.
Standard registration · Safe for all
L2 Workspace I/O 45 Skills
Reads or writes files within local project directory. Sandboxed workspace execution recommended.
Confined project boundary
L3 Network Read 47 Skills
Queries public HTTP/HTTPS documentation, search endpoints, or APIs. Requires egress monitoring for privacy.
Outbound network audit required
L4 Script Execution 277 Skills
Executes Python, Node.js, or Bash automation scripts. All helper processes verified by static AST code audits.
Subprocess review & authorization
L5 High-Privilege 31 Skills
Interacts with system credentials, tokens, or reverse-engineered services. Requires explicit administrative approval.
Mandatory administrative audit