codex-delegate-worker
claude通过私密配置的模型节点使用原生 Codex,支持 Local/Global 与原生会话
24 skills
GitHub 知识库 — 搜索仓库/Issue/PR/代码,生成仓库蓝图(架构分析+设计亮点+线稿图),中文输出
Plugin github-kb from marketplace smallraw-skills in SmallRaw/smallraw-skills. Commands read from the repo's manifest — 14 days ago.
agent-farm-delegationUse when the user explicitly invokes $agent-farm-delegation or asks to delegate a small, already-scoped task to an ACPX-backed Pi, Codex, or local ACP worker. Do not use for ordinary planning, broad architecture work, local model setup without delegation, ambiguous coding tasks, or implicit subagent
anthropomorphic-tech-writer当用户明确要直接创作一篇以技术对象拟人、第一人称自述、技术寓言、章回故事、生态兴衰史、设计推导或威胁升级为核心的原创中文技术文章时使用;也用于把事实材料重新设计成新的角色、场景和冲突。若任务要求复杂产品比较、独立多角色研究审稿或高风险事实闸门,改用 technical-story-explainer;若已有事实闭合的故事底稿且只保留原戏码润色表达,改用 technical-story-explainer-polisher。不要用于正式 API 文档、操作手册或纯文学小说。
app-analyzer用户要求分析、逆向、拆解某个 macOS 桌面应用(.app)时触发。包括但不限于:分析 App 技术栈、查看 Bundle 结构、提取资源文件、识别框架依赖、分析二进制符号。不适用于 iOS .ipa 或 Android .apk。
artifact-designUse whenever the user wants material, findings, events, processes, images, or screenshots turned into a visual artifact: technical documents, research reports, postmortems, image or screenshot comparisons, evidence boards, timelines, process maps, landing pages, slide decks, interactive prototypes,
cc-setupUser types /cc-setup to interactively configure Claude Code settings — skill usage logging hook, HUD statusLine, and future extensions.
codex-delegate-workerRun Codex through a configured custom model node with native settings and session features. Configure Local or Global nodes privately. Use when the user requests the custom worker, custom node, or its setup.
computer-useDrives the GUI of a visible local app window through `orca computer`: accessibility tree, clicks, typing, menus, dialogs, and screenshots in native apps and external browser windows (Chrome, Edge, Safari) or webviews. Prefer a programmatic path (shell, filesystem, git, HTTP, existing CLIs) whenever
excel-lite-cliExcel/xlsx/xls 报表分析、数据清洗与导出。自动处理复杂报表(表头不在首行、合并单元格、脏字符)。用户提到任何 Excel 文件的读取、查询、分析、清洗或导出时必须触发——禁止用 Read 工具直接读 xlsx(会乱码)。不修改原始文件,结果导出为新文件。
fractal-docs当用户请求初始化、更新或检查项目的模块知识文档时使用;复用 README.md 或 MODULE.md,兼容已有分形文档。
github-kbUse whenever remote GitHub content would help as evidence: the user provides an owner/repo, a github.com URL, asks to search or explore GitHub, or asks how an open-source project implements, fixes, or evolves something. Treat this as a frequently used auxiliary research source; combine it with local
guardrails-agent-pluginLoad when the prompt names `guardrails-agent-plugin`, or when a guidelines skill directs installing or updating its policy module as persistent enforcement; never infer invocation from the task alone.
guidelines-codingBehavioral guidelines to reduce common LLM coding mistakes. Use when writing, reviewing, or refactoring code to avoid overcomplication, make surgical changes, surface assumptions, and define verifiable success criteria.
guidelines-gitGit and GitHub write rules. Use before staging, committing, changing branches or history, pushing, or writing through GitHub CLI.
guidelines-security-localCheck secret access and deceptive hosts before reading/exporting credentials, private keys, auth stores, clipboard, command history, environment or crash dumps. Skip ordinary source searches and edits.
guidelines-security-npmCheck dependency acquisition, unfamiliar code execution, fetch-and-run commands and publishing across package managers. Skip routine manifest edits and trusted project scripts.
guidelines-security-shellCheck privilege escalation, permanent deletion, system permissions/disks, broad process kills, container/volume destruction and dynamic shell execution. Skip ordinary reads, edits, tests and literal wrappers.
mcp-lazy-cliUse when the user wants to invoke MCP tools, connect to MCP servers, check the mcp-registry, or manage the MCP daemon.
openclaw-tmux-agent当需要将子任务委派给其他 AI CLI 工具(claude-code、opencode、gemini-cli、codex)、并行运行多个实例、检查任务状态或断线重连恢复时触发。
orca-cliOperate Orca-managed worktrees, folder contexts, terminals, repos, automations, artifacts, skill sharing, worktree comments, and Orca's embedded browser through the `orca` CLI. Use when the user says "$orca-cli", "Orca worktree", "child worktree", "spawn codex/claude in a worktree", "read/wait/send
rule-gardener用户说 "加个规则"、遇到重复问题想固化经验、或初始化新项目规范时触发。
show-me-code当用户要求图解代码或展示调用关系、组件结构、实现差异时使用。用简洁图示帮助理解代码,仅限软件开发场景。
skill-writing-guideTRIGGER with skill-creator when creating, editing, or reviewing skills. Adds structural constraints and Anthropic's best practices — not a replacement, an add-on.
technical-story-explainer-polisher当用户已有技术科普清楚稿、故事底稿、technical-story-explainer 产物或"逻辑正确但干/不好看/说明书腔/不够沉浸"的稿子,需要润色成沉浸式中文技术故事、视频稿或长文时触发。
technical-story-explainer把复杂技术概念、产品差异、AI/API/Agent 生态或软件工程机制写成零基础读者能看懂、能复述的沉浸式中文故事。用于需要核查事实、用一个连贯故事解释底层原理,并在结尾揭示真实术语和产品差异的任务。
通过私密配置的模型节点使用原生 Codex,支持 Local/Global 与原生会话
24 skills
Codex 专属智能视觉上下文与带宽守护插件:通过滑动窗口算法自动回收历史超重 Base64 图片,杜绝多轮看图与 Computer Use 流量爆炸
24 skills
Excel 报表分析与数据清洗 - 处理复杂/乱序报表,自动清洗脏字符,查询/清洗/导出新文件
24 skills
模块知识维护 - 保存非显然的运行约束、设计原因与验证依据,复用普通 Markdown
24 skills
跨 Agent Hook 实现层 - 将领域 Skill 的规则、预检和阻断要求转换为当前 Agent 的原生 Hook
24 skills
编码行为准则 - 减少常见 AI 编码错误,并按任务阶段控制验证范围与频率
24 skills
This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.