Agent安装

Name: Agent安装
Author: 0-CODE

工程开发已审计 v1.1.0

作者仓库星标 13
作者更新于实时读取
作者仓库 viepilot

流狐档案事实层信息：领域、Agent、信任分、运行环境、权限与网络

领域: 工程开发
兼容 Agent: Claude Code
Cursor
Cline
Codex
Windsurf
Gemini CLI

+20
信任分: 92 / 100 · 已通过审计
作者 / 版本 / 许可: @0-CODE · v1.1.0 · 未声明 license
Token 消耗评级: 中等消耗
接入复杂程度: 需简单配置
是否需要外部 API Key: 不需要
兼容的系统: macOS · Linux · Windows
底层运行要求: Node.js
文件与系统权限: 只读
允许写入 / 修改
Shell 执行
网络行为: 仅限本地
安装命令数: 26 条

档案由构建时根据 SKILL.md 与安装命令自动衍生，可能与作者实际意图存在差异。

需要注意： 未限定 allowed-tools，默认拥有全部工具权限。

输出预览 vp-brainstorm.preview

---
name: vp-brainstorm
description: Brainstorm session to collect ideas and decisions for the project Output this banner as the firs…
category: 工程开发
runtime: Node.js
---

# vp-brainstorm 输出预览

## PART A: 任务判断
- 适用问题：代码实现、重构、调试或代码审查。
- 输入要求：目标材料、限制条件、期望输出和验收方式。
- 证据边界：围绕“Invocation Banner / Version Update Check (ENH-072) / Persona Context Injection (ENH-073)”读取原文规则，不把推断写成作者承诺。

## PART B: 执行结果
- **01** 任务判断：确认你的需求是否属于代码实现、重构、调试或代码审查，并标出输入、限制和预期结果。
- **02** 执行计划：优先按“Invocation Banner / Version Update Check (ENH-072) / Persona Context Injection (ENH-073)”拆成步骤，说明每一步会读取什么、修改什么、产出什么。
- **03** 交付结果：给出可复制的命令、文件改动、检查清单或内容草稿，并说明如何继续迭代。
- **04** 风险边界：结合 读取文件、写入/修改文件、执行终端命令、主要在本地完成、通常不需要额外 API Key 给出执行前确认项。

## Running Rules
- 读取文件、写入/修改文件、执行终端命令；主要在本地完成；通常不需要额外 API Key。
- 先小样例验证，再放大到真实任务。
- 交付时同时给结果、检查口径和下一步迭代建议。

SKILL.md SKILL.md

---
name: vp-brainstorm
description: Brainstorm session to collect ideas and decisions for the project Output this banner as the firs…
category: 工程开发
source: 0-CODE/viepilot
---

# vp-brainstorm

## 什么时候使用
- vp-brainstorm 是一个工程开发方向的技能，扩展 Agent 在写代码、做 review、跑测试这类场景下的能力 适合处理工程开发场景下的代码实现、调试、重构、测试或代码审查，核心价值是把输入、判断、执行、验证和交付边界固定下…
- 面向代码实现、重构、调试或代码审查，优先处理能明确输入、步骤和验收标准的工作。

## 需要提供什么
- 目标材料、目录范围、期望结果和不可改动内容。
- 是否允许联网、执行命令、读写文件或调用外部服务。

## 执行规则
- 围绕「Invocation Banner / Version Update Check (ENH-072) / Persona Context Injection (ENH-073)」组织步骤，不把推断写成作者事实。
- 读取文件、写入/修改文件、执行终端命令；主要在本地完成；通常不需要额外 API Key。
- 先跑小样例，确认结果可检查后再扩大任务范围。

## 输出要求
- 给出最终产物、关键证据、验证方式和下一步动作。
- 信息不足时标记 unknown，不编造命令、平台或依赖。

证据边界与执行链路

作者原文负责流程事实；仓库文件负责来源和命令；流狐只补充适用场景、限制和质量判断。

skill "vp-brainstorm" {
  输入层     -> 用户目标 + 目标文件 + 禁止范围 + 验收标准
  上下文层   -> Invocation Banner / Version Update Check (ENH-072) / Persona Context Injection (ENH-073)
  规则层     -> SKILL.md 触发条件 / 执行顺序 / 输出格式
  运行层     -> Node.js | 读取文件、写入/修改文件、执行终端命令 | 主要在本地完成
  安全层     -> 通常不需要额外 API Key + 小任务验证 + diff / 日志复核
  输出层     -> 可复制结果 + 检查清单 + 下一步迭代
}

解读按原文结构重写，命令、链接、术语均保留；右侧可核对作者原始 SKILL.md

先判断是否适合

核心用途：vp-brainstorm 是一个工程开发方向的技能，扩展 Agent 在写代码、做 review、跑测试这类场景下的能力适合处理工程开发场景下的代码实现、调试、重构、测试或代码审查，核心价值是把输入、判断、执行、验证和交付边界固定下来，避免 Agent…
适用任务：更适合工程开发场景下的代码实现、调试、重构、测试或代码审查，尤其是输入、步骤、验收口径可以反复复用的场景。
不宜硬套：如果任务只有一次性问答、来源材料很少，或需要作者没有说明的命令和外部服务，应先降级为普通对话。

作者设计意图

触发依据：使用时机、上下文收集、输出格式和失败处理共同决定它是否适合当前任务。
来源边界：作者原文负责事实，仓库文件负责命令和文件证据；Fluxly 只补充用途、限制和使用判断。

作者的方法与取舍

执行主线：任务围绕「Invocation Banner / Version Update Check (ENH-072) / Persona Context Injection (ENH-073)」组织，把触发条件、必要上下文、执行规则和交付标准连成一条可复核的链路。
设计重点：重点不在重复教程，而在解释作者为什么规定这些步骤、哪些条件不可省略，以及结果偏离时应回到哪一层检查。
使用判断：先确认任务与原文边界一致，再决定是否启用；来源没有说明的命令、平台和依赖保持未知，不由 Fluxly 补写。

边界和复核

依赖条件：通常不需要额外 API Key，适合先用小任务验证触发是否稳定。
权限提醒：会涉及 read / write / shell-exec 权限，执行前要说明可读写范围、命令边界和回退方式。
质量标准：结果应包含明确产物、必要证据和下一步动作；如果只是泛泛解释，需要补充输入和验收条件后重跑。

## Invocation Banner

Output this banner as the first thing on every invocation — before questions, work, or any other output:

━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
 VIEPILOT ► VP-BRAINSTORM  v1.1.0 (fw 2.19.0)
━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━

## Version Update Check (ENH-072)

After displaying the greeting banner, run:

node "$HOME/.claude/viepilot/bin/vp-tools.cjs" check-update --silent

If exit code = 1 (update available — new version printed to stdout): Display notice banner before any other output:

┌──────────────────────────────────────────────────────────────────┐
│ ✨ ViePilot {latest_version} available  (installed: {current})   │
│    npm i -g viepilot && vp-tools install --target {adapter_id}   │
└──────────────────────────────────────────────────────────────────┘

Replace {latest_version} with stdout from the command, {current} with the installed version, {adapter_id} with the active adapter (claude-code / cursor / antigravity / codex / copilot).

If exit code = 0 or command unavailable: silent, continue.

Suppression rules:

--no-update-check flag on skill invocation → skip this step entirely
config.json → update.check: false → skip this step entirely
Show at most once per session (update_check_done session guard)

## Persona Context Injection (ENH-073) At skill start, run: ```bash node "$HOME/.claude/viepilot/bin/vp-tools.cjs" persona auto-switch node "$HOME/.claude/viepilot/bin/vp-tools.cjs" persona context ``` Inject the output as `## User Persona` context before any task execution. Silent if command unavailable or errors. ## A. Skill Invocation - Skill được gọi khi user mention `vp-brainstorm`, `/vp-brainstorm`, hoặc yêu cầu "brainstorm" - Treat all user text after the skill mention as `{{VP_ARGS}}`

B. User Prompting

Prompt user conversationally with numbered list options.

C. Tool Usage

Use Claude Code tools: Bash (shell), Read (file), Edit + Write (file write/patch), Grep (search), Glob (file patterns), LS, WebSearch, WebFetch, Agent (spawn subagent — multi-level nesting supported) Interactive: AskUserQuestion (deferred — preload via ToolSearch before first call)

D. Session Actions — Proactive AUQ Triggers (BUG-026)

No slash sub-commands. All session actions are triggered proactively by the AI via AskUserQuestion at the right moment — no user typing required.

Action	Trigger condition	AUQ prompt
Save + crystallize handoff	User signals done: "save", "done", "xong", "lưu", "finished", "ready", "crystallize", or equivalent intent	"Save session and prepare for /vp-crystallize?"
UX walkthrough	In UI session AND user mentions "research", "test", "UX", "walkthrough" — OR ≥1 UI signals accumulated	"Run UX walkthrough (3-phase simulate + research + update HTML)?"
Architecture review	User asks "review decisions", "check arch", "summarize choices", or equivalent intent	"Generate architecture review summary table?"
Sync UI → Arch	After arch update when UI workspace also active	"Sync architecture changes to UI Direction workspace?"

Why: In Claude Code terminal, any /command is intercepted by the shell as a skill invocation. Using proactive AUQ triggers works identically on ALL adapters — no adapter-specific typing rules needed.

## A. Skill Invocation Same trigger keywords as claude-code adapter.

C. Tool Usage

Use Cursor tools: run_terminal_cmd (shell), read_file (read), edit_file (write/edit), grep_search (search), web_search, codebase_search, list_dir, file_search Interactive: text list fallback (AskQuestion available in Plan Mode only; Agent Mode = text) Subagent: /multitask (user command, single-level only — not a callable tool) MCP limit: 40 tools

## A. Skill Invocation Same trigger keywords as claude-code adapter. Skill discovery: LLM-driven (automatic, no slash command needed).

C. Tool Usage

Use Antigravity tools: shell (cmd), file_read, file_write, MCP plugins Interactive: text fallback (TUI-based; no formal AskUserQuestion) Skill path: .agents/skills/<skill>/SKILL.md (project) or ~/.gemini/antigravity/skills/ (global) Note: Gemini CLI deprecated June 18, 2026 — use Antigravity CLI.

## A. Skill Invocation Same trigger keywords as claude-code adapter.

C. Tool Usage

Use Codex tools: container.exec (sandboxed shell), apply_patch (file write), web_search Interactive: text fallback (TUI Tab/Enter injection) Config: ~/.codex/config.toml

## A. Skill Invocation Same trigger keywords as claude-code adapter. Discovery: User-driven (`@agent-name` in GitHub Copilot Chat).

C. Tool Usage

Use Copilot tools: runCommands (shell), read/readfile (read), edit/editFiles (write), code_search, find_references Interactive: askQuestions (main agent only — NOT available in subagents; VS Code issue #293745) Skill path: .github/agents/<name>.agent.md

ViePilot Namespace Guard (BUG-004)

Default mode: only use and reference vp-* skills in ViePilot workflows.
External skills (non vp-*) are out of framework scope unless user explicitly opts in.
If external skills appear in runtime context, ignore them and route with the closest built-in vp-* skill.

## Implementation routing guard (ENH-021)

Scoped to session artifacts (docs/brainstorm/*, .viepilot/ui-direction/*). Does not implement on behalf of /vp-auto for lib/, tests/, or framework workflows//skills/ changes — after brainstorm use /vp-crystallize / /vp-evolve → /vp-auto depending on the stage. Explicit override — state it in chat.

Collect ideas, requirements, and architecture decisions for the project through interactive Q&A.

Supports:

Create a new session
Continue a previous session
Review a past session
Landing page layout discovery (follow-up questions to finalize layout)
In-session research (research during the brainstorm session on demand)
UI Direction mode: create/update HTML prototype + notes under .viepilot/ui-direction/{session-id}/ — supports multi-page (pages/{slug}.html + hub index.html) and the ## Pages inventory hook in notes.md when pages/ exists (FEAT-007)
Phase assignment (ENH-030): in every session, features/capabilities are assigned directly to Phase 1, Phase 2, Phase 3... — no MVP/Post-MVP/Future tiers. Session file stores assignments in the ## Phases section.
Project meta intake (FEAT-009): after scope locked, before Completed / /end, if .viepilot/META.md (viepilot_profile_id) is missing — run sequential Q&A with proposals; read/write ~/.viepilot/profile-map.md; create ~/.viepilot/profiles/<slug>.md + binding per docs/dev/global-profiles.md. If a profile is already bound — skip intake by default (ask if change needed).
UX walkthrough (FEAT-010 + ENH-019 + ENH-020): in --ui mode, proactively offered via AUQ when user mentions "research", "test UX", "walkthrough" — or when ≥1 UI signals accumulated. Runs 3 phases: simulates end-user (with content stress pass + stress recipes by archetype → Stress findings) → UX designer + web research → update index.html / pages/*.html / style.css and write ## UX walkthrough log in notes.md (sync hub + Pages inventory for multi-page).
Background UI extraction (ENH-026 + ENH-060): automatically detects UI signal keywords in every brainstorm session (no --ui flag required). Auto-suggests itself (ENH-060): if the initial message contains ≥1 UI keyword, shows a proactive 🎨 banner immediately — mirrors Architect Design Mode's 🏗️ proactive activation. Accumulation starts at ≥1 signal (was ≥3); surfaces for confirmation when topic ends, user signals done, or ≥2 unique signals accumulated (was ≥5) — does not interrupt the main conversation.
Idea → Architecture breakdown loop (ENH-061): structured 8-step flow from free idea collection → scope lock → Feature → Coverage mapping (maps each Phase 1 feature to an architect page + UI screen, outputs ## Coverage matrix in notes.md) → Architect Design → arch_to_ui_sync reverse sync (surfaces UI implications of architectural decisions; proactively offered via AUQ after arch updates) → UI Direction → completeness gate (CHECK 4: warns if any Phase 1 feature has no coverage in either mode, non-blocking). See Recommended Breakdown Ordering section in workflows/brainstorm.md.
Unified workspace mode selection (BUG-018): after scope lock, before any design workspace is created, an AUQ prompt offers: Both / Architect only / UI Direction only / Neither. Architect auto-activate heuristic is deferred until after this selection; suppressed if user selects "Neither" or "UI Direction only".
Admin & Governance coverage (ENH-063): Topic 6 in brainstorm template; proactive 🔐 heuristic fires on admin/governance keywords; admin coverage gate before /save for multi-user/SaaS/compliance projects; admin.html added to Architect workspace when applicable; notes.md ## admin YAML exported via crystallize.
Content Management coverage (ENH-065): Topic 7 in brainstorm template; proactive 🗂️ heuristic fires on content-type/media/SEO keywords; content coverage gate before /save for projects with content layer; content.html added to Architect workspace; notes.md ## content YAML exported via crystallize.
Admin Entity Management coverage (ENH-068): Topic 7 in brainstorm template; proactive 🗄️ heuristic fires on CRUD/entity/admin panel keywords; entity management coverage gate before /save for projects with DB entities (cross-references ERD); entity-mgmt.html added to Architect workspace; notes.md ## entity_mgmt YAML exported via crystallize.
User Data Management coverage (ENH-066): Topic 9 in brainstorm template; proactive 👤 heuristic fires on user-account/privacy/auth keywords; user data coverage gate before /save for B2C/SaaS/GDPR projects; user-data.html added to Architect workspace; notes.md ## user_data YAML exported via crystallize.
Architect Design Mode (FEAT-011): /vp-brainstorm --architect or auto-activate when ≥3 components/services detected; generate HTML workspace (architecture, data-flow, decisions, tech-stack, tech-notes, feature-map) with Mermaid diagrams; incremental update per decision; architecture review offered via AUQ when user asks to review decisions; machine-readable notes.md YAML schema.
ERD page (ENH-027): Architect workspace includes erd.html — Mermaid erDiagram, entity list table, relationship summary; triggered by DB/entity/table/relationship keywords; notes.md ## erd YAML section exported to ARCHITECTURE.md ## Database Schema via crystallize Step 1D.
User Use Cases page (ENH-028): Architect workspace includes user-use-cases.html — actor/use-case diagram (Mermaid flowchart), use case table; triggered by user/role/actor/story keywords; notes.md ## use_cases YAML section exported to PROJECT-CONTEXT.md ## User Stories & Use Cases via crystallize Step 1D.
C4Context/Sequence/Deployment/APIs pages (ENH-029, 12-page workspace): Architect workspace expanded to 12 pages — sequence-diagram.html (per-scenario sequenceDiagram), deployment.html (infra graph + environments + CI/CD pipeline), apis.html (endpoint tables with HTTP method badges); page boundary rules table; trigger keywords for sequence/deploy/API; notes.md ## apis YAML section; deployment+APIs exported to ARCHITECTURE.md via crystallize Step 1D (sequence excluded — scenario docs are not architecture artifacts).
Embedded Domain Mode (ENH-071): Activated when ≥2 embedded trigger keywords detected (MCU families: STM32/ESP32/nRF52/AVR/RISC-V/RP2040; concepts: firmware/bare-metal/GPIO/interrupt/HAL/bootloader/RTOS) OR --domain embedded flag. Adds 6 Architect workspace pages for embedded hardware artifacts, injects domain-specific topic probes, suppresses web-UI false-positives for hardware display keywords, and offers a firmware phase ordering template. crystallize exports 8 YAML sections to ARCHITECTURE.md hardware sections.

Creates/Updates:

docs/brainstorm/session-{YYYY-MM-DD}.md

After: Ready for /vp-crystallize

Language configuration (ENH-032):

Step 0 reads ~/.viepilot/config.json → BRAINSTORM_LANG = language.document (default: en).
BRAINSTORM_LANG is used for brainstorm file storage and generated content.
User session language takes precedence over config if different.
Configure via: vp-tools config set language.document vi

Workflow version stamps (ENH-067):

Session files record workflow_version (current ViePilot semver) at create/save time.
upgrade_supplement_version stamped after gap-detection supplement completes (idempotency guard).

Upgrade gap detection (ENH-067):

On --continue, compares session workflow_version vs. current Topics Template to detect missing topics.
Proactive 🔄 upgrade banner (AUQ) lists missing topics; user can discuss inline, defer to /save, or skip.
Inline supplement appended as ## Upgrade supplement (vX → vY) section; upgrade_supplement_version stamped for idempotency.
After supplement: suggests /vp-crystallize --upgrade to patch project artifacts.

@$HOME/{envToolDir}/workflows/brainstorm.md Optional flags: - `--new` : Force create a new session - `--continue` : Continue the most recent session - `--list` : List all sessions - `--landing` : Prioritize the Landing Page layout discovery flow - `--research` : Enable proactive research suggestions during the session - `--ui` : Enable UI Direction mode (live HTML/CSS direction artifacts + auto-generates `design.md` when design keywords present — ENH-076) - `--domain embedded` : Force-activate Embedded Domain Mode (hardware topology, RTOS, pin map, memory layout, protocol matrix, power budget pages + topic probes) - `--no-stakeholders` : Skip stakeholder agent generation at session end

Reference URL Input (ENH-092)

When user provides a URL during brainstorm (competitor app, documentation, reference product):

research-agent is dispatched with op: browse_url using vercel-labs/agent-browser
Extracted insights added as ## Reference Research section in session output
Covers: features, UX patterns, pricing model, tech stack clues
Available as context for crystallize step
Falls back to WebFetch if agent-browser not installed (static HTML only — warns if page is JS-rendered)

Examples:

/vp-brainstorm I want to build a task manager. Reference: https://linear.app
/vp-brainstorm --new Compare these two tools: https://notion.so https://coda.io

Multiple URLs: dispatches compare_products op → side-by-side comparison table.

Embedded Domain Mode (ENH-071)

Activation: Automatically when ≥2 embedded keywords detected, or via --domain embedded flag. One-time 🔌 Embedded Domain Mode activated banner shown.

Topic probes injected when embedded_domain: true:

MCU/Toolchain (Gap 2): MCU family, toolchain (GCC-ARM/Keil/IAR), build system (CMake/PlatformIO/West), debug interface (SWD/JTAG), flasher, SDK/HAL
RTOS/Scheduling (Gap 3): bare-metal vs RTOS choice, task list, ISR table, resource protection strategy
Power Budget (Gap 7): supply type, sleep modes, current targets, battery life (triggered by battery/power/sleep keywords)
Safety/Compliance (Gap 10): IEC 61508/ISO 26262/DO-178C, watchdog, fault handlers (triggered by safety/SIL/ASIL keywords)
Firmware Phase Template (Gap 9): Board Bring-Up → Drivers → RTOS → Middleware → Application → Integration Test → OTA

6 new Architect workspace pages (in .viepilot/architect/{session-id}/):

Page	Trigger	Content
`hw-topology.html`	Always in embedded domain	MCU block diagram (Mermaid graph TD) + component/bus/power-rail tables
`pin-map.html`	GPIO/pin/pinout keywords	Pin assignment table (Pin# / GPIO / Function / Peripheral / Direction / Pull / Voltage)
`memory-layout.html`	Flash/RAM/linker/bootloader/OTA keywords	Flash + RAM regions tables + linker constraint notes
`protocol-matrix.html`	CAN/I2C/SPI/BLE/LoRa/MQTT/Modbus keywords	Bus protocol + wireless connectivity tables (distinct from `apis.html` HTTP REST)
`rtos-scheduler.html`	RTOS/FreeRTOS/task/ISR/scheduler keywords	Task priority table + ISR table + state diagram (≤5 tasks)
`power-budget.html`	Battery/sleep/power/µA/mAh keywords	Power modes table + battery life estimate

Pages linked in index.html under an Embedded nav section.

UI Direction false-positive suppression (Gap 6):

LCD / OLED / TFT / display / screen in hardware context → routed to hw-topology peripherals, NOT UI Direction buffer
Hardware context confirmed by: GPIO / SPI / I2C / driver / framebuffer / ILI9341 / SSD1306 / LVGL / u8g2 keywords
🎨 UI Direction Mode? banner suppressed when all display signals have hardware context

notes.md YAML sections written: ## hw_topology, ## pin_map, ## memory_layout, ## protocols, ## rtos_config, ## embedded_toolchain, ## power_budget, ## safety_config

crystallize Step 1D item 13 exports:

## Hardware Architecture (from hw_topology)
## Hardware Interface (from pin_map)
## Memory Map (from memory_layout)
## Communication Protocols (from protocols) — distinct from ## APIs
## RTOS & Task Model (from rtos_config)
## Toolchain & Build System (from embedded_toolchain)
## Power Budget (from power_budget)
## Safety & Reliability (from safety_config)
## Embedded Domain: true + MCU family written to PROJECT-CONTEXT.md
UI Coverage Gate skipped; Hardware Coverage Check (driver-task completeness, non-blocking) applied instead

Execute workflow from `@$HOME/{envToolDir}/workflows/brainstorm.md`

Key steps:

Detect existing sessions
Ask user intent (new/continue/review)
Load context if continuing
Run interactive Q&A with topic-based structure
If topic is a landing page: ask follow-up layout questions + reference 21st.dev to suggest sections/components
If topic needs UI/UX: create/update UI Direction artifacts under .viepilot/ui-direction/{session-id}/ — legacy: index.html + style.css + notes.md; multi-page: add pages/*.html, index.html as hub, and after each page change update ## Pages inventory in notes.md (see docs/user/features/ui-direction.md) 6b. When user types /research-ui or /research ui during a UI session: follow workflows/brainstorm.md exactly (FEAT-010) — do not merge into the short regular research step
If user requests research or needs to clarify a decision: research inline in the session and return to the topic
When a topic adds/modifies a capability: update ## Phases in the session per workflows/brainstorm.md
Before completing the session: step 5 — Project meta intake (FEAT-009) in workflows/brainstorm.md when binding is missing; sequential Q&A + profile-map disambiguation + write global profile + .viepilot/META.md
Save session with structured format (including ## Project meta intake (FEAT-009), research notes + UI direction references + ## Phases)
Suggest next action: /vp-crystallize

Stakeholder Generation (ENH-098): At session end (scope locked), automatically infers project stakeholders from the domain and creates .claude/agents/{project-slug}-{role}.md agent files. These are consumed by /vp-crystallize Step 1G for parallel context review. Skip with --no-stakeholders.

- [ ] Session file created/updated with all required sections - [ ] Decisions documented with rationale - [ ] Open questions listed - [ ] Action items captured - [ ] Landing page topics include explicit layout selection questions - [ ] 21st.dev references included when relevant - [ ] Research can be executed inside the same brainstorm session - [ ] UI Direction artifacts created/updated when UI mode is active - [ ] Multi-page sessions: hub links + `## Pages inventory` stay in sync with `pages/*.html` - [ ] **FEAT-010 + ENH-019 + ENH-020**: `/research-ui` (when `--ui`) runs all 3 phases, including **content stress pass** + **archetype recipes** + **`## UX walkthrough log`** (with **Stress findings**) when prototype is updated - [ ] `## Phases` present with Phase 1 having real content when scope is discussed - [ ] **FEAT-009**: intake completed, binding already present, **or** waiver with reason before Completed; session records **`## Project meta intake (FEAT-009)`** - [ ] **ENH-076**: `design.md` generated in session directory when UI mode active and ≥2 design keywords present; `notes.md ## design_tokens` populated - [ ] Next steps suggested

Adapter Compatibility

AskUserQuestion Tool (ENH-048)

This skill uses adapter-aware interactive prompts. Behavior depends on your adapter:

Adapter	Interactive Prompts	Notes
Claude Code (terminal)	✅ `AskUserQuestion` tool — REQUIRED	Must call AUQ; plain-text only if tool errors or is unavailable
Claude Code (VS Code ext)	⚠️ Partial	Terminal yes; VS Code UI pending anthropics/claude-code#12609
Cursor (Plan Mode)	⚠️ Partial	`AskQuestion` in Plan Mode only — not in Agent/Skills Mode
Cursor (Agent/Skills)	❌ Text fallback	AskQuestion not available in Agent Mode
Codex CLI	❌ Text fallback	Native tool N/A; community MCP available
Antigravity (native agent)	❌ Text fallback	Artifact model, no raw tool calls
GitHub Copilot	✅ `/skill-name` in Chat	Via `.agent.md` custom agent; AUQ not available — text fallback

Claude Code (terminal) — AUQ preload required (ENH-059): Before the first interactive prompt, call ToolSearch with query: "select:AskUserQuestion" to load the deferred tool schema. Only after ToolSearch succeeds can AskUserQuestion be invoked. If ToolSearch returns an error, fall back to plain-text numbered list for that session.

When AskUserQuestion is not available on other adapters, the skill automatically falls back to plain-text numbered list prompts — no configuration required.

Prompts using AskUserQuestion in this skill:

Session intent (continue / review / new — Step 2)
Landing page layout selection (Step 4 — Layout A/B/C/D)
Mid-session structured choices (≥2 discrete named options — BUG-022): topic direction, section priority, angle selection, and any numbered decision with enumerable options
Session-transition/next-steps prompt (session-flow choices: save/crystallize, update UI artifacts, continue discussing — BUG-023)

Skill Registry Integration (FEAT-020)

When UI Direction Mode is active, vp-brainstorm automatically:

Loads ~/.viepilot/skill-registry.json (written by vp-tools scan-skills)
Matches installed skills by capabilities to current UI signals
Silently applies matched skill best_practices to HTML generation
Records matched skills in notes.md ## skills_used

No prompt is shown — integration is transparent. Skills with relevant capabilities (e.g., ui-generation, component-design, responsive-layout) are detected automatically.

Skill decisions are locked at crystallize time (see /vp-crystallize). /vp-auto executes with those locked decisions — no re-asking.

Install skills via vp-tools install-skill <source>. See: docs/user/features/skill-registry.md

安装命令

一键复制安装选择包管理器,点击右侧复制,粘贴到终端回车即可

npx skills add 0-CODE/viepilot/skills/vp-brainstorm -y

查看通用安装教程

查看 Skill 原文

兼容工具

Claude Code
Cursor
Cline
Codex
Windsurf
Gemini
OpenCode
Roo Code
AAide
+Augment
GitHub Copilot
OpenClaw
KKiro
Junie
WWarp
Goose
Trae
Kilo Code
OpenHands
Amp
Zencoder
CContinue
AIAider
CodeBuddy
CodeGeeX
Replit Agent

安全审计

92 / 100

已审计

未声明 license/许可。
未限定 allowed-tools，默认拥有全部工具权限。

元信息

来源: github
0-CODE/viepilot
版本: 1.1.0
English

@0-CODE 的其他技能

分类与标签

分类工程开发

先判断是否适合

作者设计意图

作者的方法与取舍

边界和复核

B. User Prompting

C. Tool Usage

D. Session Actions — Proactive AUQ Triggers (BUG-026)

C. Tool Usage

C. Tool Usage

C. Tool Usage

C. Tool Usage

ViePilot Namespace Guard (BUG-004)

Reference URL Input (ENH-092)

Embedded Domain Mode (ENH-071)

Adapter Compatibility

AskUserQuestion Tool (ENH-048)

Skill Registry Integration (FEAT-020)

讨论