菜单
这个仓库中的 skills
SkillsMP 已收集 Loveacup/jz-skills 中的 81 个 Skill。打开任一 Skill 可查看来源和详情。
Loveacup/jz-skills已展示 40 / 81 个已收集 Skill。
Use when managing, evaluating, configuring, or testing text-to-speech providers for Hermes or local agent workflows. Covers provider registry, fallback policy, voice/sample tests, resource benchmarks, artifact/noise checks, and keeping TTS decisions…
原文语言：英语
Inspect codebases w/ pygount: LOC, languages, ratios.
原文语言：英语
GitHub auth setup: HTTPS tokens, SSH keys, gh CLI login.
原文语言：英语
Review PRs: diffs, inline comments via gh or REST.
原文语言：英语
Create, triage, label, assign GitHub issues via gh or REST.
原文语言：英语
GitHub PR lifecycle: branch, commit, open, CI, merge.
原文语言：英语
Clone/create/fork repos; manage remotes, releases.
原文语言：英语
GitHub 全操作入口：认证、issues、PR、仓库管理、代码审查、源码探索、代码库统计、README 写作规范。Use when working with any GitHub repository — opening issues, creating PRs, reviewing code, searching source code, cloning repos, managing CI, creating releases, counting LOC, writing or reviewing…
原文语言：英语
AI/科技领域的小红书短图文创作工作流。从选题评估、爆款拆解、风格融合到反AI味道检测，输出可直接发布的300-800字技术科普笔记。Use when: 用户要写AI/技术类小红书内容、需要科技选题评估、需要把技术话题转成小红书爆款、需要爆款笔记拆解、需要账号风格提取。DO NOT use for: 公众号长文（4000+字）、非AI/技术领域的小红书（穿搭/美妆/美食等）、爬取小红书数据（封号风险）、用户特定的飞书审稿管线。触发词：写小红书、小红书文案、XHS文案、AI技术内容、科技小红书、技术科普、AI小红书、…
Reference for Supermemory — the cloud long-term memory infrastructure for AI agents (api.supermemory.ai, NOT self-hosted/Docker). Covers concepts, SDK/API, container tags, processing pipeline, and operational troubleshooting. Two deployments here: (1) Hermes…
Use when managing cross-session AI agent memory — persistent fact storage with temporal decay, knowledge graph consolidation, deduplication, and multi-agent recall for 史馆 archiving. Based on ICM (rtk-ai/icm, 221⭐, Rust binary, MCP native) and mnem…
原文语言：英语
Use when performing independent audit and evaluation of agent execution results — verifying outputs against plans, detecting regressions, checking governance compliance, and producing structured audit reports for 御史台 review. Based on OpenAgentBench…
原文语言：英语
Use when tracking or controlling AI agent API spending — hard budget enforcement with circuit breaking, per-agent/per-task cost attribution, multi-model pricing (2,600+ models), cost optimization recommendations, and spending forecasts for 户部 budget…
原文语言：英语
三省六部金融研究 skill — 通过 Kanban 多 profile 体系安全调度 AI 金融研究，从监国承旨到史馆归档的全链路工作流。
Use when operating or reviewing the regent 三省六部 governance system: task routing (L0-L3), plan-preview triggers, decision cards, review-recovery separation, escalation rules, handoff schema v2, acceptance checks, and proactive Kanban artifact delivery.
原文语言：多语言混合
Use when setting up or debugging AI agent observability — distributed tracing across multi-agent chains, span-level cost attribution, real-time monitoring dashboards, and anomaly detection for 工部 infrastructure. Based on Laminar (lmnr-ai/lmnr, 2.9K⭐, YC S24)…
原文语言：英语
工部磁盘清理策略 — cron output 自动清理、缓存回收、日志轮转。参考 docker_cleanup.sh + homebutler 模式
工部基础设施健康巡检 — 检查 system/gateway/config/disk/process/cron 状态，参考 sudoyasir/OverWatch + cubyverse/health-checks + rz0re/spyd
工部基础设施监控 — 基于 kOlapsis/maintenant + Tracer-Cloud/opensre 模式，Docker/K8s 自动发现、HTTP/TCP/SSL 检测、心跳监控、资源指标、告警引擎
Control and troubleshoot the user's household Surge for Mac gateway via surge-cli. Use when the user asks about Surge, proxy mode/group/selection, network routing, DNS/cache, node latency, recent requests, household device identification, waking or SSH-ing…
原文语言：英语
Use when performing deep research and knowledge synthesis — multi-source web/academic search, hierarchical task graphs, evidence-backed claim verification, structured literature reviews, and long-horizon agentic research for 翰林院 knowledge work. Based on…
原文语言：英语
send_message 投送门闸 — 在 kanban 工作流中规避 check_fn 拦截的策略
Use when deploying external specialist AI agents for complex engineering tasks — domain-expert hiring pipeline, adversarial code review, multi-model routing, and external contractor integration for 将作监 operations. Based on Nexus Hyper Agent Team…
原文语言：英语
礼部 Markdown→PDF 渲染 — 基于 md-bookify(MCP) + any2pdf(CJK) + mdPress 模式，手机版/桌面版/EPUB 多格式输出
Use when performing the six-ministries smoke test — end-to-end verification of the full 三省六部 chain: 中书→门下→尚书→六部→史馆. Must run after modifying six-ministries system/config/profiles. 六部运转冒烟测试 — 端到端验证三省六部全线：中书→门下→尚书→六部→史馆。修改六部制度/配置/profile…
原文语言：多语言混合
Use when the kanban gate enforcement engine blocks a tool call or when diagnosing Kanban permission/state/cleanliness violations — the five-layer hard-intercept gate for agent tool calls. 制度引擎 — kanban_gate.py 五层硬拦截闸门，移植自 edict 的 kanban_update.py，对所有 Agent 的…
原文语言：多语言混合
Use when orchestrating multi-agent work through Kanban — decomposition playbook, anti-temptation rules, and routing patterns for an orchestrator profile. The 'don't do the work yourself' rule and the basic lifecycle are auto-injected into every kanban…
原文语言：英语
Use when working as a Kanban worker agent — pitfalls, examples, and edge cases for Hermes Kanban task execution. The lifecycle itself is auto-injected into every worker's system prompt as KANBAN_GUIDANCE (from agent/prompt_builder.py); this skill is what you…
原文语言：英语
Use when managing the AI agent fleet — capability discovery, agent registration, heartbeat monitoring, dynamic team composition, and talent pool management for 吏部 registry operations. Based on agentregistry (agentregistry-dev/agentregistry, 245⭐,…
原文语言：英语
Use when designing or implementing agent-to-agent communication in the 三省六部 system — A2A Protocol (Google→Linux Foundation, 23.4K⭐) integration for inter-agent task delegation, capability discovery, streaming artifacts, and cross-profile handoff. Based on…
原文语言：英语
Use when auditing AI agent code for security vulnerabilities — prompt injection detection, confused-deputy attacks, MCP config auditing, tool-chain exfiltration, and OWASP Agentic Top 10 compliance for 刑部 security review. Based on agent-audit…
原文语言：英语
刑部代码审查工具集 — 基于 super-linter(10.4K⭐) + roborev(1.2K⭐) + inspect(95%召回) 模式，lint/安全审计/实体级审查
中国法务综合技能，提供合同审核、法律咨询和文书起草服务。覆盖七大领域：(1)合同法务-审核商业合同/租赁协议/服务合同并提供风险分析和修改建议 (2)劳动人事-处理劳动合同/经济补偿金计算/解雇程序问题 (3)知识产权-商标注册/专利申请/版权保护咨询 (4)公司治理-股权架构设计/融资合规/章程起草 (5)民事诉讼-起诉流程指导/证据规则/执行程序 (6)个人法务-婚姻家庭/遗产继承/房产交易/消费维权…
方法论文档创作工作流。将用户的思维框架、实践经验或核心洞察，转化为结构严谨、证据充分、可传播的方法论文档。核心能力：(1) 方法论内核提取——从碎片化输入中提炼核心框架 (2) 跨领域证据矩阵构建——多维度交叉验证核心论点 (3) 结构化写作——根据素材内容自然组织文档架构 (4) 多平台适配——针对不同发布平台生成兼容版本 (5) 渐进采纳路径设计——降低读者认知门槛。触发词：写方法论、创建方法论、方法论文档、methodology、把我的经验写成方法论、系统化我的做法、总结成方法论、框架文档化。
小红书图文卡片制作工具。将文章、方法论、知识点等内容转化为适合小红书发布的精美卡片图组。核心能力：(1) 内容分析与卡片规划——从源文档提炼核心信息，规划卡片数量和内容分配 (2) Notion 风格设计系统——白底、Callout 块、Emoji 图标、大字排版 (3) HTML→PNG 精确渲染——Playwright 截图，1080×1440px（3:4 小红书标准比例）(4) 视觉 QA 闭环——渲染后检查溢出、填充率、字号，自动修复 (5) 内容审查——对照源文件验证准确性，防止信息失真 (6) PDF…
战略洞察长文工作流 v4.2 — 基于S-T-D立方体+多维分析矩阵+GoT自适应路径优化，将商业现象、行业趋势、企业战略转化为结构化深度洞察报告。 Three execution modes (default Deep): - Deep: 16 Agent + GoT adaptive + CoV anti-hallucination verification, 12-20 min - Standard: 9 core Agent, 7-10 min (user must explicitly request)…
Detect and remove AI writing patterns from prose. Bilingual engine: English (30+ patterns) and Chinese (25+ Chinese-native patterns with register-aware detection). Use when editing, rewriting, or reviewing text to eliminate predictable AI tells and inject…
性格本位的多维命运深度解析框架（v3）。立场：性格决定命运，玄学辅证，命运可塑。荣格八维为主语刻画性格签名，八字解构能量、紫微对应舞台、占星映照节律；玄学不预言，只印证、映照、解释。分析师代命主分析，输出成品命书。触发词：命运分析、命理解析、八字、紫微斗数、紫微命盘、星座分析、荣格八维、认知功能、人格分析、大运流年、合盘、夫妻宫、BaZi、cognitive functions。
Detects and resolves ambiguity in user requests — when statements are vague, contradictory, missing key constraints, or have multiple valid interpretations, grills one question at a time until precise understanding is reached. Also responds to explicit…
理解 Pi（Windows）与 Hermes（Mac Mini）联动架构、排查跨机器 MCP 连接问题。当用户问 'pi和hermes怎么连的'、'MCP 架构'、'为什么连不上hermes'、或需要新增 MCP 服务器时使用。不要用于单一 MCP 故障 — 用 recover-hindsight-mcp。