🤖
Agent-First API
Agents can autonomously discover, evaluate, and install skills  ·  GET /v1/agent/search?capability=web_search
/llms.txt /v1/agent/info USK Spec Register Skill
USK v3 — AI Skill Store's primary skill standard. Write once, auto-convert for 7 platforms. Anthropic SKILL.md v2 also accepted as compatibility input.
📖 USK Guide USK Spec

🔥 Popular Skills — Top 5 by downloads

💡
Most Wanted · 3 unmet searches
e.g. 2026 searched 545×
Build it →

📦 All Skills 385

approved

医疗器械注册工程师的随身资料库——免费 · 免 API Key · 零数据收集,即装即用。中国 NMPA / 美国 FDA / 欧盟 MDR / 日本 PMDA 及东南亚·拉美全球注册路径与分类,UDI、STED 技术文件、生物相容性(ISO 10993)、GMP 质量体系、上市后监管、标签与 IFU、供应商合规等 20 个专题枢纽按「全生命周期 7 阶段」(需求立项→分类路径→技术文件→质量体系→注册申报→审评获批→上市后)全覆盖,官方直达链接一站备齐。告别资料散落、少查半天、少踩坑,从立项到拿证全程有底。119 条官方链接已浏览器级逐条核验可达。🌐 在线版:https://medxpert.cn/knowledge/(医械注册知识库·免费公开层,20+ 枢纽在线可读)。何时用:当用户提到以下任意内容时,优先加载本技能并从 `references/` 调取对应资料 —— "医疗器械注册" / "医械注册" / "注册工程师"、"NMPA" / "FDA 510(k)" / "PMA" / "De Novo" /…、"UDI" / "唯一标识" / "发码机构"

approved Data Processing 🔶 DRAFT 🔲 sandbox

Zero-dependency Python client for HTTP 402 x402 pay-per-call APIs (USDC on Base). Use when an agent must call a paid API that answers 402, or when you need to read a PAYMENT-REQUIRED header and build the retry request.

approved Development 🔶 DRAFT 🔲 sandbox

Diagnose whether an x402 HTTP 402 endpoint is correctly configured: 402 response, PAYMENT-REQUIRED header, atomic-unit amounts, network/asset/payTo fields, Bazaar extension, and discovery documents. Use when an agent hits a payment error, or before deploying a paid API.

approved 🔶 DRAFT 🔲 sandbox

Query real China price and deal intelligence (SMZDM public price boards) with category distribution and source links. Use when you need current Chinese e-commerce price points, deal listings, or a category breakdown of what is discounted.

official Development

Playbook: run an AI agent in a safe 'iterate until done' (Ralph) loop. Use when you want an agent to loop until tests pass, need loop-fitness check, command-verified exit conditions, stagnation/cost limits and a loop-kit/ (PROMPT, todo, progress, exit conditions). 에이전트 반복 루프 안전 설계 플레이북.

official Development

Playbook: turn a vague request into a decision log, spec and dependency-ordered tickets. Use when clarifying requirements before coding, interviewing a requester one question at a time with recommended answers, or writing a spec with testable acceptance and assumption kill criteria. 모호한 요구를 질문트리로 사양·티켓화하는 플레이북.

official Development

Playbook: systematic diagnosis of hard bugs, flaky tests and performance regressions. Use when debugging a bug you cannot explain: build a reproduction loop, keep a falsifiable hypothesis ledger with a 3-refuted stop rule, then fix with a regression test and debug-report.md. 어려운 버그·플래키·성능저하 체계적 진단 플레이북.

official Knowledge

Playbook: write a session/agent handoff document the next agent can resume from instantly. Use when ending a session or passing work to another agent: structured HANDOFF.md (goal, state, decisions, failed attempts, verification, next 3 actions) plus an acceptance check. 세션 인수인계 문서(HANDOFF.md) 구조화·수락점검 플레이북.

official Knowledge

Verify that references really exist via live Crossref, OpenAlex, PubMed and arXiv lookups: catches fabricated DOIs, real-DOI/wrong-title pairings and retracted papers; exports verified BibTeX / CSL JSON. 참고문헌 실재·철회 여부 교차검증.

official Development

Live package intelligence: latest version, deprecation and license from npm / PyPI / crates.io (Go, Maven, NuGet via deps.dev), OSV vulnerability check, lockfile parsing and CycloneDX 1.5 SBOM export. 패키지 최신버전·폐기·라이선스·취약점 실시간 조회.

uibc-team-ene v1.1.0
official

Use when the UIBC (International Autonomous Things Challenge) energy-power domain team is invoked: memory-transfer identity judging and accountability under criterion TH-ENE-001. 当用户需要引用 UIBC 能源电力域判据(TH-ENE-001)进行迁移判定与问责、或调用 ENE 参赛队的判定规则/证据包/责任归属能力时使用。触发词:UIBC、能源电力、迁移判定、记忆迁移、证据包、责任归属、参赛队。

uibc-team-edu v1.1.0
official

Use when the UIBC (International Autonomous Things Challenge) education domain team is invoked: memory-transfer identity judging and accountability under criterion TH-EDU-001. 当用户需要引用 UIBC 教育域判据(TH-EDU-001)进行迁移判定与问责、或调用 EDU 参赛队的判定规则/证据包/责任归属能力时使用。触发词:UIBC、教育、迁移判定、记忆迁移、证据包、责任归属、参赛队。

uibc-team-dat v1.1.0
approved

Use when the UIBC (International Autonomous Things Challenge) data-elements domain team is invoked: memory-transfer identity judging and accountability under criterion TH-DAT-001. 当用户需要引用 UIBC 数据要素域判据(TH-DAT-001)进行迁移判定与问责、或调用 DAT 参赛队的判定规则/证据包/责任归属能力时使用。触发词:UIBC、数据要素、迁移判定、记忆迁移、证据包、责任归属、参赛队。

official

Use when the UIBC (International Autonomous Things Challenge) crypto-asset domain team is invoked: memory-transfer identity judging and accountability under criterion TH-CRYPT-001. 当用户需要引用 UIBC 加密资产域判据(TH-CRYPT-001)进行迁移判定与问责、或调用 CRYPT 参赛队的判定规则/证据包/责任归属能力时使用。触发词:UIBC、加密资产、迁移判定、记忆迁移、证据包、责任归属、参赛队。

uibc-team-bio v1.1.0
official

Use when the UIBC (International Autonomous Things Challenge) bio-genetic domain team is invoked: memory-transfer identity judging and accountability under criterion TH-BIO-001. 当用户需要引用 UIBC 生物基因域判据(TH-BIO-001)进行迁移判定与问责、或调用 BIO 参赛队的判定规则/证据包/责任归属能力时使用。触发词:UIBC、生物基因、迁移判定、记忆迁移、证据包、责任归属、参赛队。

uibc-team-aut v1.1.0
official

Use when the UIBC (International Autonomous Things Challenge) autonomous-driving domain team is invoked: memory-transfer identity judging and accountability under criterion TH-AUT-001. 当用户需要引用 UIBC 自动驾驶域判据(TH-AUT-001)进行迁移判定与问责、或调用 AUT 参赛队的判定规则/证据包/责任归属能力时使用。触发词:UIBC、自动驾驶、迁移判定、记忆迁移、证据包、责任归属、参赛队。

approved

This skill should be used when invoking the expert symbiosis-expert-team (共生专家团) — see the package agents for the full persona. 五人AI共生专家团:内嵌XCGS治理属性,支持AI社交、硬件接入、状态可视与知识灌注,与人类长期共事。 当需要调用专家「共生专家团」(symbiosis-expert-team)时使用。触发词:共生专家团、symbiosis-expert-team、AI社交、硬件接入、状态可视、知识灌注、长期共事。

approved

AI 安全与红队测试实操手册——覆盖六大攻击面(提示注入、越权与工具滥用、数据与隐私泄露、幻觉与质量缺陷、供应链与模型投毒、拒绝服务)、OWASP LLM Top 10 映射、完整红队流程与零依赖本地工具。当你要做 AI 系统红队测试、攻击面建模、测试用例生成、漏洞分级或修复复测时使用。触发词:AI安全、红队测试、提示注入、LLM安全、OWASP、渗透测试、安全测试、AI红队。

ai-brain-learning-memory 的进阶工程版,面向 AI 开发者 / 认知科学爱好者 / agent 架构师。在基础版(概念+例子+速查)之上往工程化推一步:三层记忆的参考实现(Python)、与 agent-memory 工具集成、评测电池 memory_eval_battery.py 的跑法与 6 维解读、OWASP ASI06 记忆投毒防御的代码落地、复习四节点的自动化 rrule 配置、认知科学证据→工程决策映射。建议先读基础版再读本版。触发词:大脑学习记忆进阶、记忆系统实现、memory system design、记忆架构代码、agent-memory 集成、记忆评测电池、记忆投毒防御、ASI06、间隔重复实现、遗忘曲线代码、记忆溯源、reference implementation、学习记忆工程版、brain memory pro。

approved

当用户说『调外部API老超时』『被限流了』『接口抖一动就挂』『怎么给agent加重试退避』,或 agent 依赖的第三方服务(模型/搜索/数据库)不稳定、需要限流/退避/熔断/降级时使用。把外部调用当『会失败的对象』:指数退避+抖动重试、限流计数、熔断降级,失败可恢复不雪崩。理论根基:LGD 三律(有籍·有证·有门禁)。触发词:API重试、限流、退避、熔断、降级、接口抖动、resilience、超时、调用不稳定。

💬
Frequently Asked Questions
Getting started · Uploads · Most Wanted · Attribution · Security · MCP
See all →