harness-cli

Route AIOS host dispositions and execute the current rex-harness software Capability Command. TRIGGER: 分析、设计、实现、调试、并发、agent team、长任务、harness、plan、工作流、多步骤

AIOS-native long-running agent harness with rex Command execution, ContextDB, checkpoint recovery, and evidence capture. Use for a planned resumable objective when AIOS is installed. If AIOS is NOT installed, use `harness-init-runner` for a lightweight standalone alternative.

Initialize a lightweight repo-local Node.js harness (harness/ + .harness/) WITHOUT AIOS dependency. Use ONLY when you need a standalone, portable harness. If AIOS is installed, use `aios-long-running-harness` instead — it has rex Command hosting, ContextDB integration, and checkpoint recovery.

Quick reference for code-review-graph (CRG) MCP tools. Use when you need to look up a CRG tool name, parameter, or pattern. For workflow guidance, see AGENTS.md decision checkpoints. Requires `aios internal codemap install`.

RTK + Caveman community token compression tools install and config guide. RTK (github.com/rtk-ai/rtk) compresses command output 60-90%. Caveman (github.com/JuliusBrussee/caveman) cuts output tokens ~75%. Both run locally, no external services. The AIOS native interception runtime is deprecated.

Use when recalling prior AIOS tool/browser outputs from offloaded refs; inspect Mermaid canvas first, then read only matching node-level evidence.

Use when operating in the aios repository and needing the canonical architecture, memory schema, MCP browser tool behavior, and execution constraints before changing workflows.

Create a DESIGN.md style baseline BEFORE building UI. Use FIRST when no design draft exists — installs a proven visual style, then invoke `frontend-design` to implement. For borrowing a known product style, getting consistent typography/colors/spacing, or needing a fast visual starting point.

Use when the user sends `cap` or asks for a fast commit-and-push shortcut for current repository changes.

Automatic ContextDB session lifecycle — init, session events, checkpoints, and continuity artifacts. Use when you need AIOS context persistence without prompt injection. NOT for general task execution — only for ContextDB session management. TRIGGER: contextdb、session persist、checkpoint save.

Evidence-first debugging WITH MCP tools. Use when debug_hub MCP server is available. Inject log calls, collect via HTTP API (port 39200), analyze through MCP tools, clean up when done. If debug_hub MCP is unavailable, use `debug` skill (bundled Python collector) instead.

Helps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express interest in extending capabilities. This skill should be used when the user is looking for functionality that might exist as an installable skill.

Build production-grade frontend UI with distinctive design. Use AFTER establishing design direction — if no DESIGN.md exists, invoke `awesome-design-md` first to create one, then use this skill to implement. For web components, pages, landing pages, dashboards, React components, HTML/CSS layouts, or styling any web UI.

当需要调度不同模型执行子任务时使用。根据任务类型匹配模型能力，自动选择最优模型并生成调用指令。TRIGGER: 模型调度、model dispatch、选模型、分派任务、多模型协作、路由到模型

Use before ANY code modification — editing files, creating new files, or changing behavior. Mandatory CRG impact-radius check, dependency tracing, style alignment, and test-coverage verification before edits. Post-edit CRG update and verification.

Use when the user wants to generate an image, edit an image, create visual artwork, produce text-to-image output, transform an existing image with image-to-image, or asks for 生图, 生成图片, 文生图, 图生图, text-to-image, image-to-image, AI image, picture generation, image result, image edit, or visual generation.

Research-before-building workflow. Use when adding new behavior, dependencies, or integrations to avoid reinventing and to stay compatible across codex/claude/gemini/opencode.

Lightweight security hygiene for agent configs (skills/hooks/MCP/settings). Use before enabling automation, after config changes, and before release.

Use when improving a single page for SEO and AI answer-engine visibility (GEO), especially when you need faster first-screen conversion and clearer traffic funnels.

Enforce browser automation safety constraints — tool usage order, anti-detection rules, screenshot discipline, and human-in-the-loop checks. Use when setting up or verifying browser/MCP tool operations, NOT for every skill execution.

Train and optimize skill documents using the SkillOpt methodology — the agent acts as both target and optimizer, running tasks, reflecting on failures, proposing edits, and validating improvements in a self-contained training loop. TRIGGER: optimize skill, train skill, improve skill, skill training, skill optimization, SkillOpt, skill evolution, skill iteration loop, skill feedback loop, skill tuning.

Guide for creating effective skills. This skill should be used when users want to create a new skill (or update an existing skill) that extends the agent's capabilities with specialized knowledge, workflows, or tool integrations.

Evidence-before-assertions workflow. Use before claiming work is done, before release, and after any behavior change in scripts/skills/MCP.

Use when completing a task and deciding whether repository changes require a semantic version bump and changelog entry before commit/push.

Install Codex skills into $CODEX_HOME/skills from a curated list or a GitHub repo path. Use when a user asks to list installable skills, install a curated skill, or install a skill from another repo (including private repos).

Create new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, update or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.