viepilot 2.15.0 → 2.23.0

package/CHANGELOG.md

@@ -7,6 +7,143 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [Unreleased]
 
+## [2.23.0] - 2026-04-18
+
+### Added
+- **ENH-059**: AskUserQuestion ToolSearch preload — all vp-* skills and workflow files on Claude Code adapter now instruct the model to call `ToolSearch { query: "select:AskUserQuestion" }` before the first interactive prompt, preventing deferred-tool fallback to text menus
+  - 5 SKILL.md files updated: `vp-request`, `vp-evolve`, `vp-auto`, `vp-brainstorm`, `vp-crystallize`
+  - 3 workflow files updated: `workflows/request.md`, `workflows/evolve.md`, `workflows/autonomous.md`
+  - `vp-auto/SKILL.md` now has a full AUQ adapter compatibility section
+
+## [2.22.1] - 2026-04-18
+
+### Fixed
+- **BUG-015**: `copilot` adapter missing from `bin/viepilot.cjs` CLI — `TARGETS` array and `printHelp()` text now include `copilot`; `viepilot --list-targets` shows GitHub Copilot; `viepilot install --target copilot` no longer fails with "Unknown target"
+
+## [2.22.0] - 2026-04-18
+
+### Added
+- **ENH-058**: Workflow continuation prompt — AskUserQuestion at vp-evolve and vp-request completion
+  - `workflows/evolve.md` Step 5: AUQ asks "Execute now / Create request / Done" — selecting Execute invokes `/vp-auto` immediately
+  - `workflows/request.md` Step 7: AUQ asks "Plan phase / Create another / Done" — selecting Plan invokes `/vp-evolve` immediately
+  - Text fallback for Cursor / Codex / Copilot / Antigravity adapters (static list preserved)
+  - `skills/vp-evolve/SKILL.md` + `skills/vp-request/SKILL.md` AUQ prompt tables updated
+
+## [2.21.0] - 2026-04-18
+
+### Added
+- **FEAT-019: GitHub Copilot Adapter** — install and run vp-* skills inside VS Code Copilot Chat and GitHub Copilot CLI
+  - `lib/adapters/copilot.cjs` — adapter definition with `~/.config/gh-copilot/` config home
+  - Auto-detect via `~/.config/gh-copilot/` directory or `gh` CLI binary presence
+  - Skill invocation: `/vp-status`, `/vp-auto`, etc. (same slash syntax as Claude Code / Cursor)
+  - Install command: `viepilot install --target copilot`
+  - 4 adapter-table SKILL.md files updated with GitHub Copilot row
+  - `docs/user/features/adapters.md` — Copilot section with surface matrix, prerequisites, limitations
+
+## [2.20.0] - 2026-04-18
+
+### Added
+- **ENH-057: ViePilot Agents System** — 6 dedicated sub-agents for repetitive/parallelizable skill tasks:
+  - `tracker-agent`: TRACKER.md read/write delegation (phase status, task status, decision log)
+  - `research-agent`: WebSearch + WebFetch feasibility studies (auto-triggered in request.md Step 2B)
+  - `file-scanner-agent`: Glob + Grep repo-wide scanning (Explore subagent on Claude Code)
+  - `changelog-agent`: atomic CHANGELOG + version bump — single authority (resolves ENH-053)
+  - `test-generator-agent`: contract test scaffolding + run from acceptance criteria
+  - `doc-sync-agent`: bulk multi-file `.md` updates (≥5 files → 1 agent call)
+- `agents/` directory with 6 agent definition files (`agents/*-agent.md`)
+- Research-agent feasibility gate in `workflows/request.md` Step 2B — auto-triggered for Feature/platform requests
+- Bulk-edit task detection in `workflows/autonomous.md` — doc-sync-agent invoked when ≥5 identical file types in task Paths
+- Agent delegation table + invoke-agent patterns in `workflows/autonomous.md`
+- `docs/dev/agents.md` — developer reference for agents layer architecture and invocation
+
+### Fixed
+- **ENH-053**: Version bump authority unified in `changelog-agent` — `autonomous.md` and `evolve.md` both invoke it; no more inline duplication
+
+## [2.19.0] - 2026-04-18
+
+### Added
+- **ENH-056: Skill invocation greeting banner** — all 17 `vp-*` skills now output
+  a version banner as the very first output on invocation:
+  ```
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+   VIEPILOT ► VP-AUTO  v0.2.2 (fw 2.19.0)
+  ━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+  ```
+  Addresses Claude Code UI change where skill-load indicators are no longer shown;
+  users can now confirm which skill version is running. 119 contract tests added.
+
+## [2.18.0] - 2026-04-18
+
+### Fixed
+- **vp-rollback** (BUG-014): `rollback.md` Step 7 now parses enriched tag format
+  `{project}-{branch}-{version}-vp-p{N}-t{M}` introduced by ENH-050. Rollback to
+  any Phase 57+ checkpoint correctly restores HANDOFF.json phase/task state.
+  All 3 formats (legacy, project-scoped, enriched) handled via `grep -oE vp-p[0-9]+`.
+
+### Changed
+- **vp-crystallize** (ENH-051): Added "Brownfield Execution Path" table clarifying
+  which sub-steps (1A–1D) run, skip, or are conditional in brownfield mode.
+  Steps 1A/1B CONDITIONAL; Steps 1C/1D SKIP with rationale.
+- **vp-brainstorm** (ENH-052): Step 6 now validates phase assignment before saving.
+  Scope-locked sessions with features but no phase assignments are blocked with
+  actionable message. Exploratory sessions and brownfield stubs bypass the gate.
+- **Version bump rules** (ENH-053): Consolidated into `.viepilot/SYSTEM-RULES.md`
+  as single canonical table with precedence rule (MAJOR > MINOR > PATCH) and
+  mixed-phase handling. `evolve.md` and `autonomous.md` reference this table
+  instead of defining rules inline.
+- **vp-audit post-phase hook** (ENH-054): Auto-hook integration upgraded from
+  conceptual note to concrete `<step name="post_phase_audit">` block in
+  `autonomous.md`. Runs 3 fast Tier 1+2 checks after every phase-complete event;
+  completely silent when no issues found; y/n prompt when issues detected.
+- **AskUserQuestion enforcement** (ENH-055): All 13 AUQ prompt blocks across 4
+  workflows (evolve, request, brainstorm, crystallize) now explicitly mark Claude
+  Code adapter as REQUIRED. Plain-text menus remain as fallback for non-Claude-Code
+  adapters only. `vp-evolve/SKILL.md` gains new Adapter Compatibility section.
+
+### Files
+- `workflows/rollback.md` — Step 7: 3-format tag parse with `grep -oE`
+- `workflows/crystallize.md` — Brownfield Execution Path table (Step 0-C → Step 1)
+- `workflows/brainstorm.md` — Pre-save phase validation gate in Step 6
+- `.viepilot/SYSTEM-RULES.md` — Canonical Version Bump Rules table
+- `workflows/evolve.md` — Version bump ref → SYSTEM-RULES.md; AUQ REQUIRED
+- `workflows/autonomous.md` — Version bump ref + `post_phase_audit` step (5c)
+- `workflows/audit.md` — Auto-hook: concrete integration spec replaces conceptual note
+- `workflows/request.md`, `brainstorm.md`, `crystallize.md` — AUQ REQUIRED
+- `skills/vp-evolve/SKILL.md` — New Adapter Compatibility section
+- `skills/vp-request`, `vp-brainstorm`, `vp-crystallize` SKILL.md — AUQ REQUIRED
+- `tests/unit/vp-workflow-consistency.test.js` — 35 contract tests
+
+## [2.17.0] - 2026-04-17
+
+### Changed
+- **vp-auto** git tags now include active branch and package version for full
+  traceability. New format: `{prefix}-{branch}-{version}-vp-p{n}-t{n}`. (ENH-050)
+  - Branch resolved via `git rev-parse --abbrev-ref HEAD` (sanitized: `/` → `-`)
+  - Version resolved from `package.json` (fallback: `0.0.0`)
+  - Legacy tags (`{prefix}-vp-p{n}-*`) continue to be recognized by audit + rollback
+
+### Files
+- `workflows/autonomous.md` — `TAG_PREFIX` resolution block (BRANCH_SAFE + VERSION);
+  task start, task done, and phase complete tag patterns updated to enriched format
+- `workflows/audit.md` — `COMPLETE_TAGS` regex: added third alternative for enriched
+  format; `PREV_TAG` grep: extended char class to include `.` for version strings
+- `workflows/rollback.md` — tag grep char class extended to include `.` for version strings
+- `tests/unit/vp-enh050-git-tag-format.test.js` — 16 contract tests
+
+## [2.16.0] - 2026-04-17
+
+### Changed
+- **vp-audit** Tier 4 (Framework Integrity) now runs silently — output is suppressed
+  when the check passes (✅) or is skipped (non-framework repo). Tier 4 is only
+  surfaced in the audit report when issues (⚠️) are found. (ENH-049)
+
+### Files
+- `workflows/audit.md` — Step 4 skip: removed `echo "→ Tier 4 skipped"` message;
+  Step 4f report block wrapped in `TIER4_ISSUES > 0` guard; All Clear banner Tier 4
+  line removed; Issues Found banner Tier 4 line rendered only when `TIER4_ISSUES > 0`
+- `skills/vp-audit/SKILL.md` — documented silent-by-default Tier 4 behavior (ENH-049)
+- `tests/unit/vp-enh049-audit-tier4-silent.test.js` — 13 contract tests
+
 ## [2.15.0] - 2026-04-13
 
 ### Changed (ENH-048 — Phase 78: AskUserQuestion Adapter-Aware Integration)

package/README.md

@@ -2,15 +2,15 @@
 
 **Autonomous Vibe Coding Framework / Bộ khung phát triển tự động có kiểm soát**
 
-[![Version](https://img.shields.io/badge/version-2.15.0-blue.svg)](CHANGELOG.md)
+[![Version](https://img.shields.io/badge/version-2.23.0-blue.svg)](CHANGELOG.md)
 [![License: MIT](https://img.shields.io/badge/License-MIT-green.svg)](LICENSE)
 [![Skills](https://img.shields.io/badge/skills-17-purple.svg)](#skills-reference)
 [![Workflows](https://img.shields.io/badge/workflows-13-orange.svg)](#workflows)
 [![Templates](https://img.shields.io/badge/templates-17-cyan.svg)](#templates)
-[![Tests](https://img.shields.io/badge/tests-957%20passing-brightgreen.svg)](tests/)
+[![Tests](https://img.shields.io/badge/tests-1140%20passing-brightgreen.svg)](tests/)
 [![GitHub](https://img.shields.io/github/stars/0-CODE/viepilot?style=social)](https://github.com/0-CODE/viepilot)
 
-**Versioning:** Shield **2.15.0** is the **ViePilot framework SemVer** tracked in `.viepilot/TRACKER.md` and `CHANGELOG.md`. The npm `package.json` field `version` (**2.15.0**) is the Node package identifier for this repo; use the framework version for milestone releases and docs.
+**Versioning:** Shield **2.19.0** is the **ViePilot framework SemVer** tracked in `.viepilot/TRACKER.md` and `CHANGELOG.md`. The npm `package.json` field `version` (**2.19.0**) is the Node package identifier for this repo; use the framework version for milestone releases and docs.
 
 ViePilot là bộ skill framework cho phép AI assistant (Claude, GPT, etc.) phát triển dự án một cách **tự động**, **có kiểm soát**, và **có thể khôi phục**. Thiết kế theo các tiêu chuẩn chuyên nghiệp: Semantic Versioning, Conventional Commits, Keep a Changelog.
 

package/bin/viepilot.cjs

@@ -23,6 +23,7 @@ const TARGETS = [
   { id: 'cursor-ide',   label: 'Cursor IDE' },
   { id: 'antigravity',  label: adapterMap['antigravity'].name },
   { id: 'codex',        label: adapterMap['codex'].name },
+  { id: 'copilot',      label: adapterMap['copilot'].name },
 ];
 
 function printHelp() {
@@ -36,14 +37,14 @@ Usage:
   viepilot --list-targets
 
 Install options:
-  --target <id|id,id|all>   Target profile(s): claude-code (mirrors vp-* to ~/.claude/skills), cursor-agent, cursor-ide, antigravity, codex
+  --target <id|id,id|all>   Target profile(s): claude-code, cursor-agent, cursor-ide, antigravity, codex, copilot
   --yes                      Non-interactive mode (skip confirmations)
   --dry-run                  Print actions only (Node installer; no bash)
   --list-targets             Print supported targets and exit
   --help                     Show help
 
 Uninstall options:
-  --target <id|id,id|all>   Remove assets (claude-code: ~/.claude/skills/vp-*; cursor-*: ~/.cursor/skills/vp-*; antigravity: ~/.antigravity/skills/vp-*; codex: ~/.codex/skills/vp-*)
+  --target <id|id,id|all>   Remove assets (claude-code: ~/.claude/skills/vp-*; cursor-*: ~/.cursor/skills/vp-*; antigravity: ~/.antigravity/skills/vp-*; codex: ~/.codex/skills/vp-*; copilot: ~/.config/gh-copilot/skills/vp-*)
   --yes                     Non-interactive mode (skip confirmations)
   --dry-run                 Print actions only, do not remove files
 `);

package/docs/dev/agents.md

@@ -0,0 +1,131 @@
+# ViePilot Agents Layer
+
+Developer reference for the agents system introduced in ENH-057 (v2.20.0).
+
+## Why Agents?
+
+Before v2.20.0, vp-* skills handled everything inline: research web fetches, TRACKER.md
+updates, CHANGELOG writes, file scans, test generation, and doc updates all ran
+sequentially in the main LLM context. This caused:
+
+1. **Context bloat** — long skill runs exhausted context with intermediate data
+2. **Sequential bottleneck** — research + write + verify ran one-at-a-time
+3. **Skill coupling** — each skill re-implemented the same patterns (tracker update, version bump)
+4. **No parallelism** — vp-auto tasks that could run concurrently didn't
+
+The agents layer delegates these operations to dedicated sub-agents — isolated contexts
+that run in parallel or in sequence without polluting the calling skill's conversation.
+
+## Architecture
+
+```
+vp-* skill (orchestrator)
+    │
+    ├── tracker-agent      ← .viepilot/TRACKER.md read/write
+    ├── research-agent     ← WebSearch + WebFetch (isolated context)
+    ├── file-scanner-agent ← Glob + Grep (repo-wide scanning)
+    ├── changelog-agent    ← CHANGELOG + version bump (single authority)
+    ├── test-generator-agent ← tests/*.test.js generation + run
+    └── doc-sync-agent     ← skills/*/SKILL.md bulk update
+```
+
+Agent definitions live in `agents/` at the repo root. They are installed to
+`{viepilotDir}/agents/` alongside workflows and templates.
+
+## Agent Catalog
+
+### tracker-agent
+**File**: `agents/tracker-agent.md`
+**Purpose**: Read/write TRACKER.md — phase status, task status, decision log, request table.
+**When to invoke**: vp-auto (task start/complete/phase complete), vp-request (Step 5), vp-evolve (phase creation).
+**Claude Code subagent_type**: `general-purpose`
+
+### research-agent
+**File**: `agents/research-agent.md`
+**Purpose**: WebSearch + WebFetch + summarize for feasibility studies and tech research.
+**When to invoke**: vp-request (Step 2B — auto-triggered for Feature/platform requests), vp-brainstorm (external validation).
+**Claude Code subagent_type**: `general-purpose`
+
+### file-scanner-agent
+**File**: `agents/file-scanner-agent.md`
+**Purpose**: Glob + Grep across repo to find affected files, detect stale refs.
+**When to invoke**: vp-audit (Tier 1–4), vp-evolve (impact analysis), vp-rollback (state restore).
+**Claude Code subagent_type**: `Explore` (specialized for codebase scanning)
+
+### changelog-agent
+**File**: `agents/changelog-agent.md`
+**Purpose**: Atomically append CHANGELOG entry + bump package.json version.
+**When to invoke**: vp-auto post-phase (last task), vp-evolve (milestone ship).
+**Claude Code subagent_type**: `general-purpose`
+**Note**: Single authority for version bumps — resolves ENH-053.
+
+### test-generator-agent
+**File**: `agents/test-generator-agent.md`
+**Purpose**: Generate contract tests from acceptance criteria, run suite, report pass/fail.
+**When to invoke**: Last task of each phase (the `N.last` contract-test task pattern).
+**Claude Code subagent_type**: `general-purpose`
+
+### doc-sync-agent
+**File**: `agents/doc-sync-agent.md`
+**Purpose**: Bulk-apply the same change to ≥5 `.md` files (adapter rows, banners, etc.).
+**When to invoke**: Auto-triggered by autonomous.md when task Paths block has ≥5 identical file types.
+**Claude Code subagent_type**: `general-purpose`
+
+## Invocation Patterns
+
+### Claude Code (terminal)
+
+Use the `Agent` tool with the appropriate `subagent_type`:
+
+```js
+Agent({
+  subagent_type: "general-purpose",  // or "Explore" for file-scanner
+  description: "{agent-name}: {operation-summary}",
+  prompt: `
+    Load agents/{agent-name}.md for the full specification.
+    Execute operation: {operation}
+    Inputs: {inputs}
+  `
+})
+```
+
+### Cursor / Codex / Antigravity
+
+Execute the equivalent operation inline in the same session. The agent `.md` file
+describes the exact steps — follow them as a scoped sub-prompt.
+
+## Adapter Behavior Table
+
+| Agent | Claude Code | Cursor | Codex | Antigravity |
+|-------|------------|--------|-------|-------------|
+| tracker-agent | Subagent (general-purpose) | Inline | Inline | Inline |
+| research-agent | Subagent (general-purpose) | Inline if web available | Inline if web available | Inline if web available |
+| file-scanner-agent | Subagent (Explore) | Inline | Inline | Inline |
+| changelog-agent | Subagent (general-purpose) | Inline | Inline | Inline |
+| test-generator-agent | Subagent (general-purpose) | Inline | Inline | Inline |
+| doc-sync-agent | Subagent (general-purpose) | Sequential inline | Sequential inline | Sequential inline |
+
+## Adding a New Agent
+
+1. Create `agents/{name}-agent.md` with all required sections:
+   - `## Purpose`
+   - `## Inputs`
+   - `## Outputs`
+   - `## Invocation Pattern` (Claude Code + text fallback)
+   - `## Adapter Behavior` (table covering all 4 adapters)
+   - `## Notes`
+
+2. Add the agent to the delegation table in `workflows/autonomous.md`
+   under `## Agent Delegation`.
+
+3. Wire invocation into the relevant workflow steps.
+
+4. Add to `installSubdirs` in the adapter that should include `agents/`.
+
+5. Write a contract test in `tests/unit/` verifying the file exists and has required sections.
+
+## Naming Conventions
+
+- File: `agents/{name}-agent.md` (kebab-case, always ends in `-agent`)
+- Reference: `{name}-agent` (e.g., `tracker-agent`, not `tracker`)
+- Invocation: `Load agents/{name}-agent.md for full spec.`

package/docs/user/features/adapters.md

@@ -10,6 +10,7 @@ ViePilot supports multiple AI coding platforms via its adapter system (FEAT-013)
 | `cursor-agent` / `cursor-ide` | Cursor | `~/.cursor/skills/` | `~/.cursor/viepilot/` | — | `/vp-status` |
 | `antigravity` | Google Antigravity | `~/.antigravity/skills/` | `~/.antigravity/viepilot/` | — | `/vp-status` |
 | `codex` | OpenAI Codex CLI | `~/.codex/skills/` | `~/.codex/viepilot/` | — | `$vp-status` |
+| `copilot` | GitHub Copilot | `~/.config/gh-copilot/skills/` | `~/.config/gh-copilot/viepilot/` | — | `/vp-status` |
 
 > **Note — Codex invocation syntax:** OpenAI Codex CLI uses `$skill-name` to invoke skills (e.g. `$vp-status`, `$vp-brainstorm`). The `/command` prefix is reserved for Codex built-in controls (`/plan`, `/clear`, `/diff`, etc.). SKILL.md file format is fully compatible — no changes needed to skill content.
 
@@ -28,6 +29,9 @@ viepilot install --target antigravity
 # OpenAI Codex CLI
 viepilot install --target codex
 
+# GitHub Copilot
+viepilot install --target copilot
+
 # Multiple targets at once
 viepilot install --target claude-code,antigravity
 ```
@@ -45,6 +49,53 @@ At install time, `{envToolDir}` is replaced with each adapter's `executionContex
 - `cursor` → `.cursor/viepilot`
 - `antigravity` → `.antigravity/viepilot`
 - `codex` → `.codex/viepilot`
+- `copilot` → `.config/gh-copilot/viepilot`
+
+## GitHub Copilot
+
+GitHub Copilot adapter (FEAT-019) enables vp-* skills to run inside VS Code Copilot Chat and the GitHub Copilot CLI.
+
+### Surface support
+
+| Surface | Status | Notes |
+|---------|--------|-------|
+| VS Code Copilot Chat | ✅ Supported | Invoke with `/vp-status`, `/vp-auto`, etc. |
+| GitHub Copilot CLI | ✅ Supported | Invoke with `/vp-status` in CLI chat |
+| Copilot Cloud Agents | ⚠️ Preview | GitHub Copilot Extensions (Public Preview); subject to change |
+| JetBrains Copilot | ⚠️ Pending | Copilot Chat for JetBrains — skills not yet validated |
+
+### Prerequisites
+
+- **gh CLI** installed (`brew install gh` or [cli.github.com](https://cli.github.com))
+- **GitHub Copilot subscription** (Individual, Business, or Enterprise)
+- **GitHub Copilot extension** for VS Code (marketplace ID: `GitHub.copilot-chat`)
+
+### Installation
+
+```bash
+npx viepilot install --target copilot
+```
+
+Files are installed to:
+- Skills: `~/.config/gh-copilot/skills/`
+- Workflows + lib: `~/.config/gh-copilot/viepilot/`
+
+After install, open VS Code, open Copilot Chat (`Ctrl+Shift+I` / `Cmd+Shift+I`), and type `/vp-status`.
+
+### Availability detection
+
+The installer detects Copilot by checking (in order):
+
+1. `~/.config/gh-copilot/` directory exists (primary)
+2. `gh` binary found at `/usr/local/bin/gh`, `/opt/homebrew/bin/gh`, or `~/.local/bin/gh` (secondary)
+
+### Known limitations
+
+| Feature | Status |
+|---------|--------|
+| `AskUserQuestion` interactive prompts | ❌ Not available — skills use text-based menus |
+| Hooks (PreToolUse, Stop, etc.) | ❌ Copilot uses `.agent.md` convention, not programmatic hooks |
+| MCP server integration | 🔜 Planned for Phase 2 |
 
 ## Adding a new adapter
 

package/lib/adapters/antigravity.cjs

@@ -24,7 +24,8 @@ module.exports = {
     path.join('templates', 'architect'),
     'bin',
     'lib',
-    'ui-components'
+    'ui-components',
+    'agents'
   ],
   isAvailable: (home) => {
     const h = home || os.homedir();

package/lib/adapters/claude-code.cjs

@@ -32,7 +32,8 @@ module.exports = {
     path.join('templates', 'architect'),
     'bin',
     'lib',
-    'ui-components'
+    'ui-components',
+    'agents'
   ],
   // Detection: is this platform available on the current machine?
   isAvailable: (home) => {

package/lib/adapters/codex.cjs

@@ -25,7 +25,8 @@ module.exports = {
     path.join('templates', 'architect'),
     'bin',
     'lib',
-    'ui-components'
+    'ui-components',
+    'agents'
   ],
   isAvailable: (home) => {
     const h = home || os.homedir();

package/lib/adapters/copilot.cjs

@@ -0,0 +1,44 @@
+'use strict';
+const os = require('os');
+const path = require('path');
+const fs = require('fs');
+
+module.exports = {
+  id: 'copilot',
+  name: 'GitHub Copilot',
+  // Paths (home-relative; resolved at install time)
+  skillsDir:   (home) => path.join(home, '.config', 'gh-copilot', 'skills'),
+  viepilotDir: (home) => path.join(home, '.config', 'gh-copilot', 'viepilot'),
+  // {envToolDir} in SKILL.md files resolves to this value at install time (ENH-035)
+  executionContextBase: '.config/gh-copilot/viepilot',
+  // Copilot Chat uses /skill-name (same as Claude Code / Cursor)
+  postInstallHint: 'Open Copilot Chat in VS Code and type /vp-status to get started',
+  hooks: {
+    configFile: null,  // Copilot uses .agent.md convention, not programmatic hooks
+    schema: 'copilot',
+    supportedEvents: []
+  },
+  installSubdirs: [
+    'workflows',
+    path.join('templates', 'project'),
+    path.join('templates', 'phase'),
+    path.join('templates', 'architect'),
+    'bin',
+    'lib',
+    'ui-components',
+    'agents'
+  ],
+  isAvailable: (home) => {
+    const h = home || os.homedir();
+    // Primary: gh-copilot config dir exists
+    if (fs.existsSync(path.join(h, '.config', 'gh-copilot'))) return true;
+    // Secondary: gh CLI installed at common paths (user likely has Copilot access)
+    const ghPaths = [
+      '/usr/local/bin/gh',
+      '/opt/homebrew/bin/gh',
+      '/usr/bin/gh',
+      path.join(h, '.local', 'bin', 'gh'),
+    ];
+    return ghPaths.some((p) => fs.existsSync(p));
+  },
+};

package/lib/adapters/cursor.cjs

@@ -23,7 +23,8 @@ module.exports = {
     path.join('templates', 'architect'),
     'bin',
     'lib',
-    'ui-components'
+    'ui-components',
+    'agents'
   ],
   isAvailable: (home) => {
     const h = home || os.homedir();

package/lib/adapters/index.cjs

@@ -6,6 +6,7 @@ const adapters = {
   'cursor-ide':   require('./cursor.cjs'),      // alias
   'antigravity':  require('./antigravity.cjs'),
   'codex':        require('./codex.cjs'),
+  'copilot':      require('./copilot.cjs'),
 };
 
 /**

package/lib/viepilot-install.cjs

@@ -191,6 +191,15 @@ function buildInstallPlan(packageRoot, envSource = process.env, opts = {}) {
         steps.push(ent.isDirectory() ? { kind: 'copy_dir', from: src, to: dest } : { kind: 'copy_file', from: src, to: dest });
       }
     }
+    // ENH-057: copy agents/ directory (agents layer added in v2.20.0)
+    const agentsRoot = path.join(root, 'agents');
+    if (fs.existsSync(agentsRoot)) {
+      for (const ent of listDirEntries(root, 'agents')) {
+        const src = path.join(root, 'agents', ent.name);
+        const dest = path.join(vpDir, 'agents', ent.name);
+        steps.push(ent.isDirectory() ? { kind: 'copy_dir', from: src, to: dest } : { kind: 'copy_file', from: src, to: dest });
+      }
+    }
     // BUG-007: copy package.json so resolveViepilotPackageRoot() finds the root
     steps.push({ kind: 'copy_file', from: path.join(root, 'package.json'), to: path.join(vpDir, 'package.json') });
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "viepilot",
-  "version": "2.15.0",
+  "version": "2.23.0",
   "description": "**Autonomous Vibe Coding Framework / Bộ khung phát triển tự động có kiểm soát**",
   "main": "index.js",
   "bin": {

package/skills/vp-audit/SKILL.md

@@ -4,6 +4,18 @@ description: "Audit state, docs drift, and stack best-practice compliance — wo
 version: 0.3.2
 ---
 
+<greeting>
+## Invocation Banner
+
+Output this banner as the **first** thing on every invocation — before questions, work, or any other output:
+
+```
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+ VIEPILOT ► VP-AUDIT  v0.3.2 (fw 2.19.0)
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+```
+</greeting>
+
 <cursor_skill_adapter>
 ## A. Skill Invocation
 - Skill được gọi khi user mention `vp-audit`, `/vp-audit`, "audit", "kiểm tra", "check docs"
@@ -74,6 +86,9 @@ When auditing a project bootstrapped via `vp-crystallize --brownfield`:
 - `ARCHITECTURE.md` counts vs actual `skills/`, `workflows/`, CLI
 - `README.md` viepilot-specific badges (version, skills-N, workflows-N)
 - `docs/skills-reference.md` sections vs `skills/` directory
+- **Silent by default (ENH-049):** Tier 4 output is suppressed when all checks pass (✅)
+  or when the check is skipped (non-framework repo). Output only appears when issues (⚠️)
+  are found. Non-framework repos: Tier 4 skipped silently with no message.
 
 **Output:**
 - Report by 4 tiers, each tier with its own status

package/skills/vp-auto/SKILL.md

@@ -4,6 +4,18 @@ description: "Autonomous execution loop with control points and recovery"
 version: 0.2.2
 ---
 
+<greeting>
+## Invocation Banner
+
+Output this banner as the **first** thing on every invocation — before questions, work, or any other output:
+
+```
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+ VIEPILOT ► VP-AUTO  v0.2.2 (fw 2.19.0)
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+```
+</greeting>
+
 <cursor_skill_adapter>
 ## A. Skill Invocation
 - Skill được gọi khi user mention `vp-auto`, `/vp-auto`, "auto", "vibe", "chạy tự động"
@@ -230,6 +242,22 @@ At any control point, offer:
 Display progress summary on stop.
 </process>
 
+## Adapter Compatibility
+
+### AskUserQuestion Tool (ENH-059)
+Control-point prompts use `AskUserQuestion` on Claude Code (terminal).
+
+| Adapter | Interactive Prompts | Notes |
+|---------|---------------------|-------|
+| Claude Code (terminal) | ✅ `AskUserQuestion` — **REQUIRED** at control points | Preload schema via ToolSearch first |
+| Cursor / Codex / Copilot / Antigravity | ❌ Text fallback | Plain numbered list |
+
+**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.
+
+**Prompts using AskUserQuestion in this skill:**
+- Control point decisions (retry / skip / rollback / stop — Step 7)
+
 <success_criteria>
 - [ ] Phases executed in dependency order
 - [ ] Tasks tracked with git tags

package/skills/vp-brainstorm/SKILL.md

@@ -4,6 +4,18 @@ description: "Brainstorm session to collect ideas and decisions for the project"
 version: 1.1.0
 ---
 
+<greeting>
+## 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)
+━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━━
+```
+</greeting>
+
 <cursor_skill_adapter>
 ## A. Skill Invocation
 - Skill được gọi khi user mention `vp-brainstorm`, `/vp-brainstorm`, hoặc yêu cầu "brainstorm"
@@ -115,14 +127,18 @@ This skill uses adapter-aware interactive prompts. Behavior depends on your adap
 
 | Adapter | Interactive Prompts | Notes |
 |---------|---------------------|-------|
-| Claude Code (terminal) | ✅ `AskUserQuestion` tool | Click-to-select UI, multi-select, preview panels |
+| 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](https://github.com/anthropics/claude-code/issues/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, the skill automatically falls back to
+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:**