FairladyZ625

Coordinate Harness Anything work from task creation through direct or delegated execution, durable supervision, evidence-based acceptance, integration, and task closeout. Use when acting as the accountable lead for one or more Harness tasks.

Contribute a public change to Harness Anything from a GitHub issue through an isolated worktree, scoped tests, manifest-selected gates, a complete PR body, review triage, and maintainer merge. Use when a human or coding agent is preparing or updating a…

引导用户从首次下载到开始使用 Harness Anything：检查环境、获取持久源码、安装依赖、把整套技能链接到所选 Agent 的用户目录，随后继续目标仓库初始化和主控技能首次使用。用于用户要求下载、安装全套、首次接入或一段话带完 Harness；当前使用源码分发，不假定已有 npm 发布包。

在 macOS 或 Windows 的 DeepSeek Harness 桌面端安装 Harness Anything、全部使用技能与 DSH 插件，并验证 ha、预设和教程入口。用于用户要求同时安装 Harness 与 DeepSeek Harness 插件；已有 Harness 时复用，不重新安装。

File a public issue in FairladyZ625/harness-anything from a finding — decide whether it is public, collect command and version evidence, fill the canonical source-triaged body, redact private data, and publish with `gh issue create --body-file`. Use when a…

Codex ends its turn on the first assistant message with no tool call, and an unattended dispatch never sends another user message. Use when a Codex Agent must carry progress and checkpoints through to a settlement action instead of ending on a chat summary.

Start Harness Anything task creation by choosing a software/coding preset first. Use when creating or planning a harness task package.

Create or update a Harness Anything business Agent — a declared identity with its own instructions, prompts, skills, role, preset, runtime targets, and permission mode — instead of only naming a model. Use when a user asks to create an agent, define a new…

操作 HAiKnow 团队飞书知识库与 OKR 总控台：按十进制分类放资料、更新类目索引、拆解 KR 为里程碑和任务、填写进展。往 HAiKnow 知识库加文档或链接、整理目录、更新 OKR、拆任务、写进展、看 OKR 看板时使用。

得力飞书（Lark）操作入口，三机分工：Company 电脑拉取、本机 MacBook 与 Air 只查询。用于检索/导出/归档得力飞书云文档、知识库、会议纪要与逐字稿，以及消费端按资料包还原入库。Agent 使用前必须先判定当前是哪台机器。

查询 haiknow 飞书租户（bcnncgqa2y6v.feishu.cn）的知识库、云文档、表格与多维表格。用于读取「人机协作空间」「赚钱大计」「实操项目变流量」等知识库内容，按标题检索、展开目录树、导出文档正文。当用户提到 haiknow、飞书知识库、wiki、项目看板、团队会议记录时使用。

ZeYu AI Brain 知识库标准操作规程（SOP）v2.0 所有 Agent 操作知识库的唯一权威指南。 包含：目录治理、计划文档治理、写入模式、检索规则、obsidian-cli 规范、 外部文章资产治理、Task/Todo 宪章、巡检标准。 **v2.0 核心变更**： - 废弃 Writer-Agent 中转模式，改为 Agent 直接写知识库 + 自己 commit - 废弃旧脚本（create-task-skeleton / update-task-status / archive-task） -…

快速理解泽宇的 Brain（~/Brain）：它是什么、目录怎么分、谁在写它、自动化在哪、和 Hermes 的边界、三台机器怎么同步。第一次进 Brain、要改 Brain 的结构/脚本/定时任务、排查 Brain 自动化或同步问题时使用。只找内容用 zeyu-brain-search，写入规则看 zeyu-ai-brain-sop。

Teaches agents how to search and navigate ZeYu-AI-Brain with direct search, Obsidian CLI, Agent Graph, and the Brain directory map. Use when an agent needs to find prior notes, project context, decisions, todos, people, documents, or accumulated knowledge in…

给泽宇产出单文件 HTML 报告:四种形态——比喻式、技术式、产品方向/竞品、向上汇报。向上汇报含手机优先适配与 PDF 交付。

Run read-only governance audits for ZeYu-AI-Brain knowledge base quality and compliance. Use when checking 00-09 folder standards, task skeleton completeness, active-vs-done consistency, "done but not archived" drift, missing required fields, and cron/repo…

Use when defining or packaging a custom Agora Nomos/Harness Pack through an interview-driven flow, especially when a user wants to create a shareable project-harness asset instead of only using the built-in default.

Agora task bootstrap skill. Use when an agent enters a newly created Agora task session and needs the minimum operating context: what Agora is, who the controller is, what role this agent is bound to, what stage the task is currently in, which actions are…

Use when delegating focused coding, review, or repository questions through ACPX in Agora, especially when you need explicit agent selection, safer Claude model routing, named session handling, or Agora-specific wrapper defaults.

Coding Agent Harness 工程方法论。为使用 Coding Agent（Codex、Claude Code、Gemini CLI 等） 做长程项目开发的团队，在用户的项目上构建一套完整的 harness 工程体系。 包括：项目诊断、AGENTS.md + CLAUDE.md 入口文件生成、coding-agent-harness/ 目录搭建、Planning Loop、SSoT 治理、 Delivery Operating Model、Repository…

Use when creating, reviewing, improving, or installing Coding Agent Harness preset packages for repeatable task families.

统一检索和使用 Claude Code / Codex / AntiGravity 的历史对话。支持按日期、关键词、会话 ID、Conversation Reference、Conversation Prompt 找回原文，也支持在用户明确要求时，对指定会话或某天会话做提炼、总结、笔记整理。

Coordinate complex coding and knowledge-work tasks inside a user's target repository with Codex's current main agent and in-task internal subagents. Use for multi-agent task graphs, parallel exploration, disjoint implementation slices, long-running…

面向 AI for Science 领域的高严谨中文学术报告撰写、De-AI 去油打磨、GB/T 7714-2015 文献引证与图文契约规范。

把一个模糊的项目目标,通过和用户**一个问题一个问题、无限循环地持续讨论**(对抗式追问、反思、主动澄清),逐渐逼到清晰,并渐进沉淀成一套分层、自包含、任何人读完即可开工的设计上下文包。当用户描述了一个还不完全清晰的目标——"我想做一个 X""我们重写 / 重新设计 / 改造一下 Y""帮我想清楚这个东西到底要做什么"——务必使用本 skill。它适用于全新项目和老项目改造 alike。即使用户没明说,只要目标还模糊、需要持续讨论才能清晰,就主动触发它,不要等用户说"我要做设计文档"。