opencode-vibe-spec-plugin 1.0.4

package/README.md

@@ -0,0 +1,264 @@
+# OpenCode Vibe-Spec Plugin
+
+[![Version](https://img.shields.io/badge/version-1.0.4-blue.svg)](https://github.com/blank-1/opencode-vibe-spec-plugin)
+[![License](https://img.shields.io/badge/license-MIT-green.svg)](LICENSE)
+[![Node.js](https://img.shields.io/badge/node-%3E%3D18.0.0-brightgreen.svg)](https://nodejs.org)
+
+Multi-agent plugin for OpenCode with **3 primary agent modes**: **Vibe** (rapid prototyping), **Spec** (spec-driven development), and **Ask** (research assistant) — backed by 11 specialized subagents and an integrated quality gate workflow.
+
+## Quick Start
+
+Add to your OpenCode config (`~/.config/opencode/opencode.json`):
+
+```json
+{
+  "$schema": "https://opencode.ai/config.json",
+  "plugin": [
+    "opencode-vibe-spec-plugin"
+  ]
+}
+```
+
+OpenCode will automatically install the plugin from npm on first launch.
+
+## Features
+
+| Mode | Description | Key Capability |
+|------|-------------|----------------|
+| **Vibe** | Rapid prototyping agent | Creative exploration + quick implementation |
+| **Spec** | Spec-driven development agent | EARS requirements → phase-gated workflow → implementation |
+| **Ask** | Multi-function research assistant | Codebase analysis + context enhancement |
+
+### Spec-Driven Development (New)
+
+The Spec agent now provides enterprise-grade spec management:
+
+- **EARS/GEARS Notation**: Structured requirements using `WHEN [trigger] THE SYSTEM SHALL [response]` format
+- **Phase-Gated Workflow**: `requirements → design → tasks → implementation` with validation at each gate
+- **Task Traceability**: Requirements-to-tasks linkage with full traceability matrix
+- **Persistent Spec Storage**: Per-feature specs organized under `.opencode/specs/`
+- **Steering Integration**: Auto-loads `product.md`, `tech.md`, `structure.md` context
+- **Spec Lifecycle Hooks**: `onSpecCreated`, `onPhaseAdvance`, `onSpecCompleted`
+
+### Quality Gate (Workflow Mechanism)
+
+Quality gates are integrated as **automatic workflow validation**, not a standalone agent:
+
+- **Phase Gates**: Automatic quality validation at each spec phase transition
+- **Requirements Gate**: Min 3 EARS requirements + acceptance criteria
+- **Design Gate**: Architecture + technology stack validation
+- **Tasks Gate**: Task count + requirement traceability
+
+### Additional Features
+
+- **Tab Key Switching**: Seamless mode switching with context preservation
+- **11 Specialized Subagents**: Purpose-built for different tasks
+- **Smart Model Routing**: Automatic model selection based on task type
+- **Steering Documents**: Project conventions automatically followed
+- **Todo Enforcement**: Mandatory/advisory task completion strategies
+
+## Agent Modes
+
+### Vibe Mode
+Press `Tab` → Select Vibe → Describe your goal → Get executable prototype
+
+**Styles**: Rapid Prototype | Inspiration Capture | Exploratory Coding
+
+### Spec Mode
+Press `Tab` → Select Spec → Describe requirements → Get structured specification
+
+**Formats**: EARS/GEARS Notation | Kiro-Style | Phase-Gated Workflow
+
+**Workflow**:
+```
+requirements → [Quality Gate] → design → [Quality Gate] → tasks → [Quality Gate] → implementation
+```
+
+### Ask Mode
+Press `Tab` → Select Ask → Ask questions → Get research-backed answers
+
+**Capabilities**: Code Research | Context Enhancement | Requirements Clarification
+
+## Subagents
+
+### Shared (7)
+- **Librarian**: Documentation research specialist
+- **Oracle**: Code analysis specialist
+- **Architect**: System design specialist
+- **QualityAnalyzer**: Real-time quality analysis
+- **QualityReviewer**: Quality validation
+- **QualityMetrics**: Metrics collection
+- **QualityReporter**: Report generation
+
+### Vibe (2)
+- **Inspirer**: Creativity and ideation
+- **Explorer**: Feasibility and risk assessment
+
+### Spec (2)
+- **Planner**: Requirements analysis
+- **Reviewer**: Specification validation
+
+## Architecture Patterns
+
+This plugin implements industry-leading agent architecture patterns:
+
+| Pattern | Use Case | Benefit |
+|---------|----------|---------|
+| **REWOO** | Ask Agent | 5x token efficiency vs ReAct |
+| **Hierarchical** | Spec Agent | Structured delegation |
+| **Plan-and-Solve** | Spec Agent | Sequential task execution |
+| **Reflector** | Vibe Agent | Self-improving loops |
+
+## Configuration
+
+### Model Configuration
+
+```json
+{
+  "agent": {
+    "vibe": { "model": "zhipu/glm-4-flash" },
+    "spec": { "model": "zhipu/glm-5.2" },
+    "ask": { "model": "zhipu/glm-4-flash" }
+  }
+}
+```
+
+### Environment Variables
+
+```bash
+export VIBE_SPEC_MODEL_VIBE="zhipu/glm-4-flash"
+export VIBE_SPEC_MODEL_SPEC="zhipu/glm-5.2"
+export VIBE_SPEC_MODEL_ASK="zhipu/glm-4-flash"
+```
+
+### Steering Documents
+
+Create `.opencode/steering.md` in your project root to define project conventions.
+For spec-driven development, place guidance files in `steering/`:
+
+```
+steering/
+  product.md       # Business context and goals
+  tech.md          # Technology stack and constraints
+  structure.md     # File organization patterns
+```
+
+### Spec Storage
+
+Specs are persisted under `.opencode/specs/`:
+
+```
+.opencode/
+  specs/
+    user-authentication/
+      requirements.md    # EARS-formatted requirements
+      design.md          # Technical design document
+      tasks.md           # Implementation tasks with traceability
+      metadata.json      # Spec metadata (phase, version, etc.)
+```
+
+## Quality Assessment
+
+### Quality Dimensions
+
+| Agent | Dimensions |
+|-------|------------|
+| **Vibe** | Code Readability, Technical Feasibility, Innovation, Completeness |
+| **Spec** | EARS Compliance, Structure, Executability, Traceability, Completeness |
+| **Ask** | Accuracy, Completeness, Relevance, Actionability |
+
+### Phase Gate Validation
+
+| Transition | Requirements |
+|------------|-------------|
+| requirements → design | Min 3 EARS requirements + acceptance criteria |
+| design → tasks | Architecture diagram + technology stack |
+| tasks → implementation | Min 1 task + requirement traceability |
+
+### Gate Decisions
+
+- **PROMOTE** (Score ≥ 8/10): Output ready for use
+- **HOLD** (Score 6-7/10): Needs improvements
+- **ROLLBACK** (Score < 6/10): Requires significant rework
+
+## Hooks
+
+| Hook | Purpose |
+|------|---------|
+| `tool.execute.before` | Pre-execution tool validation |
+| `message.updated` | Incremental message processing |
+| `message.completed` | Quality assessment triggers |
+| `stop.requested` | Graceful shutdown handling |
+| `session.start` | Session initialization |
+| `session.end` | Session cleanup and persistence |
+| `spec.created` | Spec file initialization |
+| `spec.phase-advance` | Phase gate validation |
+| `spec.completed` | Implementation plan generation |
+
+## Development
+
+> **Full specification**: See [`docs/spec.md`](docs/spec.md) for system architecture, core features, and success criteria.
+> **Sample steering file**: See [`examples/steering/default.md`](examples/steering/default.md) as a template for your project conventions.
+
+```bash
+# Install dependencies
+npm install
+
+# Build the plugin
+npm run build
+
+# Run tests
+npm test
+
+# Type check
+npm run typecheck
+```
+
+### From Source (Development)
+
+1. Clone the repository:
+   ```bash
+   git clone https://github.com/blank-1/opencode-vibe-spec-plugin.git
+   cd opencode-vibe-spec-plugin
+   ```
+
+2. Install and build:
+   ```bash
+   npm install
+   npm run build
+   ```
+
+3. Add to OpenCode config using local path:
+   ```json
+   {
+     "plugin": [
+       "file:///absolute/path/to/opencode-vibe-spec-plugin/dist/index.js"
+     ]
+   }
+   ```
+
+## Testing
+
+```bash
+# Run all tests (21 files, 283 tests)
+npm test
+
+# Run specific test suite
+npm test -- tests/spec/
+npm test -- tests/real-env/
+
+# Run with coverage
+npm test -- --coverage
+```
+
+## Contributing
+
+1. Fork the repository
+2. Create a feature branch (`git checkout -b feature/amazing-feature`)
+3. Commit changes (`git commit -m 'feat: add amazing feature'`)
+4. Push to branch (`git push origin feature/amazing-feature`)
+5. Open a Pull Request
+
+## License
+
+MIT License - see [LICENSE](LICENSE) for details.

package/dist/agents/ask.d.ts

@@ -0,0 +1,4 @@
+import type { AgentConfig } from '../config/types';
+export declare const askAgentConfig: AgentConfig;
+export declare const askAgentPrompt = "You are ASK - a multi-function research assistant specializing in codebase understanding, context enhancement, and requirements clarification.\n\n## CORE CAPABILITIES\n\n### 1. Code Research\n- Answer questions about codebase architecture\n- Explain how functions/modules work\n- Find documentation and usage examples\n- Analyze dependencies and relationships\n\n### 2. Architecture Pattern Analysis\n- Analyze coding workflows using REWOO patterns\n- Identify optimal architecture for different tasks\n- Map dependencies and system interrelationships\n- Detect technical constraints and opportunities\n\n### 3. Context Enhancement\n- Gather relevant context for Vibe agent (inspiration, feasibility)\n- Gather relevant context for Spec agent (requirements, architecture)\n- Provide background information for implementation decisions\n\n### 4. Requirements Clarification\n- Help clarify ambiguous user intent with structure\n- Ask targeted questions to understand requirements\n- Generate structured requirements from dialogue\n\n## WORKFLOW\n\n### Step 1: REWOO: Intelligence Planning\nBased on the user's query, analyze using REWOO pattern:\n\n**Gather (W)** - Collect necessary information:\n- If task involves investigation/exploration: use librarian, oracle, architect\n- If task requires context: use librarian, oracle, explorer\n- If task is ambiguous: use inspirer, planner\n\n**Plan (R) - Intelligent Planning**:\n1. Determine if exploration should precede coding\n2. Identify key information needs\n3. Plan information synthesis workflow\n\n**Execute (O) - Parallel Execution**:\n- Immediately call relevant subagents in parallel\n- Give each subagent 30 seconds to execute\n- Prioritize exploration tasks if they exist\n\n**Optimize (O) - Quality Loop**:\n- If initial context is insufficient, ask clarifying questions\n- Gather additional context as needed\n- Synthesize comprehensive response\n\n### Step 2: Refine Intelligence Flow\n1. **Exploration Analysis**: Examine all gathered information\n2. **Gap Detection**: Identify missing or unclear information\n3. **Assumptions Chaining**: Connect pieces into coherent understanding\n4. **Recommendation Prioritization**: Sort by impact and urgency\n\n### Step 3: Intelligent Classification\nClassify intent using architecture-aware approach:\n\n**RESEARCH** - Questions about existing code:\n- \"How does X work?\" - Investigation task\n- \"What does this function do?\" - Analysis task\n- \"Explain this architecture\" - Pattern analysis\n- \"Find documentation for Y\" - Information gathering\n\n**CONTEXT** - Providing background:\n- \"Gather context for implementing X\" - Enhancement task\n- \"What should Vibe know about this?\" - Preparation task\n- \"Prepare context for the Spec agent\" - Architecture context\n\n**CLARIFY** - Ambiguous requirements:\n- \"I want to build something like X\" - Planning assistance\n- \"Help me understand what you need\" - Requirements gathering\n- \"Let's clarify the requirements\" - Need analysis\n\n### Step 4: Synthesize Intelligence\n- Wait for ALL subagent results\n- Combine findings from multiple sources\n- Apply REWOO optimization to resolve conflicts\n- Prioritize information by relevance and impact\n\n### Step 5: Generate Response\nFormat the response with clear structure and source attribution.\n\n## RESPONSE FORMAT\n\nAlways use this structured format:\n\n```markdown\n# \uD83D\uDD0D Ask Response\n\n## Intent Classification\n**Type**: [RESEARCH | CONTEXT | CLARIFY]\n**Confidence**: [High | Medium | Low]\n\n## Intelligence Overview\n**Key Findings**: [2-3 sentence summary]\n**Architecture Pattern**: [How this relates to workflows]\n**Next Steps**: [Recommended actions]\n\n## Details\n### Relevant Files\n- `path/to/file.ts` - [What's relevant]\n- `path/to/another.ts` - [What's relevant]\n\n### Key Code Snippets\n```typescript\n// Relevant code with context\n```\n\n### Architecture Impact\n- [How this relates to overall architecture]\n- [Dependencies and relationships]\n\n## Source Attribution\n- **librarian**: [What librarian found]\n- **oracle**: [What oracle found]\n- **architect**: [What architect found]\n\n## Follow-up Suggestions\n1. [Suggested action 1]\n2. [Suggested action 2]\n\n## Confidence Score\n**Overall**: [High/Medium/Low]\n**Reasoning**: [Why this confidence level]\n```\n\n## KEY RULES\n\n1. **Always cite sources** - Never present information without attribution\n2. **Never invent file paths** - Only use paths from subagent results\n3. **Apply REWOO intelligence** - Always consider exploration before coding\n4. **Ask for clarification** - If the query is ambiguous, ask questions\n5. **Suggest follow-ups** - Help the user understand next steps\n\n## SPECIAL SCENARIOS\n\n### When you don't know\nIf you cannot find the answer:\n1. Clearly state what you searched for\n2. Explain why you couldn't find it\n3. Suggest alternative approaches\n4. Offer to help with related questions\n\n### When results conflict\nIf subagents provide conflicting information:\n1. Present both perspectives\n2. Explain the source of conflict\n3. Provide your assessment based on evidence\n4. Let the user decide\n\n### When the query is too broad\nIf the query is too vague:\n1. Ask clarifying questions\n2. Provide a structured way to narrow down\n3. Offer specific subtopics to explore\n\n## INTEGRATION WITH OTHER AGENTS\n\n### For Vibe Agent\nWhen providing context for Vibe:\n- Focus on inspiration and creativity\n- Provide feasibility assessment through REWOO patterns\n- Highlight potential approaches with architecture awareness\n- Include examples from similar projects with impact analysis\n\n### For Spec Agent\nWhen providing context for Spec:\n- Focus on requirements and architecture patterns\n- Provide dependency analysis with REWOO intelligence\n- Highlight technical constraints with workflow considerations\n- Include existing patterns and standards with optimization insights\n\n### For Build Mode\nWhen preparing for implementation:\n- Focus on implementation details with parallel execution plans\n- Provide code examples with REWOO planning\n- Highlight potential pitfalls with exploration-first approach\n- Include testing considerations with workflow optimization\n\n## EXAMPLES\n\n### Example 1: Code Research (With REWOO)\n**User**: \"How does the authentication system work?\"\n\n**Response**:\n```markdown\n# \uD83D\uDD0D Ask Response\n\n## Intent Classification\n**Type**: RESEARCH\n**Confidence**: High\n\n## Intelligence Overview\n**Key Findings**: The authentication system uses JWT tokens with refresh token rotation across AuthService, AuthMiddleware, and TokenStore components.\n**Architecture Pattern**: Linear flow with validation middleware\n**Next Steps**: Review integration patterns for new modules\n\n## Details\n### Relevant Files\n- `src/services/auth.ts` - Core authentication logic\n- `src/middleware/auth.ts` - Request validation\n- `src/stores/token.ts` - Token persistence\n\n### Key Code Snippets\n```typescript\n// Token generation\nexport function generateTokens(user: User) {\n  const accessToken = jwt.sign({ userId: user.id }, SECRET);\n  const refreshToken = crypto.randomBytes(40).toString('hex');\n  return { accessToken, refreshToken };\n}\n```\n\n## Source Attribution\n- **librarian**: Found JWT documentation and patterns\n- **oracle**: Analyzed token flow and security\n- **architect**: Mapped authentication dependencies\n\n## Follow-up Suggestions\n1. Review `src/services/auth.ts` for implementation details\n2. Check `tests/auth.test.ts` for usage examples\n3. Examine token refresh flow in `src/middleware/auth.ts`\n```\n\n### Example 2: Context Enhancement (With REWOO)\n**User**: \"Gather context for implementing a new API endpoint\"\n\n**Response**:\n```markdown\n# \uD83D\uDD0D Ask Response\n\n## Intent Classification\n**Type**: CONTEXT\n**Confidence**: High\n\n## Intelligence Overview\n**Key Findings**: The project uses Express with TypeScript, follows RESTful patterns, and has existing middleware for auth, validation, and error handling. Exploration preceded implementation analysis.\n**Architecture Pattern**: Modular API structure with middleware separation\n**Next Steps**: Consider using existing route patterns for consistency\n\n## Details\n### Existing Patterns\n- Route handlers in `src/routes/`\n- Middleware in `src/middleware/`\n- Services in `src/services/`\n- Types in `src/types/`\n\n### Recommended Approach\n1. Create route in `src/routes/`\n2. Add middleware for auth/validation\n3. Implement service logic\n4. Add tests\n\n## Source Attribution\n- **librarian**: Found Express.js documentation and patterns\n- **oracle**: Analyzed route architecture\n- **explorer**: Assessed implementation complexity\n```\n\nRemember: Your goal is to provide accurate, well-sourced information that helps the user understand their codebase and make informed decisions.";
+//# sourceMappingURL=ask.d.ts.map

package/dist/agents/ask.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"ask.d.ts","sourceRoot":"","sources":["../../src/agents/ask.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAC;AAEnD,eAAO,MAAM,cAAc,EAAE,WAmB5B,CAAC;AAEF,eAAO,MAAM,cAAc,msRAsQqH,CAAC"}

package/dist/agents/quality-gate.d.ts

@@ -0,0 +1,55 @@
+/**
+ * Quality Gate Utility Module
+ *
+ * This module provides quality gate functionality as a workflow mechanism,
+ * not as a standalone agent. It integrates with the spec workflow to
+ * validate quality at each phase transition.
+ */
+import type { GateDecision } from '../config/types';
+/**
+ * Quality dimensions for spec workflow validation
+ */
+export declare const SPEC_QUALITY_DIMENSIONS: {
+    readonly requirements: {
+        readonly earsFormat: "EARS格式合规性";
+        readonly acceptanceCriteria: "验收标准完整性";
+        readonly traceability: "需求可追溯性";
+        readonly completeness: "需求完整性";
+    };
+    readonly design: {
+        readonly architecture: "架构设计合理性";
+        readonly technologyStack: "技术栈选择";
+        readonly errorHandling: "错误处理策略";
+        readonly scalability: "可扩展性考虑";
+    };
+    readonly tasks: {
+        readonly actionability: "任务可执行性";
+        readonly dependencyOrder: "依赖顺序合理性";
+        readonly requirementCoverage: "需求覆盖度";
+        readonly granularity: "任务粒度适中";
+    };
+};
+/**
+ * Validate phase transition quality
+ */
+export declare function validatePhaseTransition(from: string, to: string, context: {
+    requirements?: string;
+    design?: string;
+    tasks?: string;
+}): {
+    valid: boolean;
+    issues: string[];
+    score: number;
+};
+/**
+ * Generate quality gate decision
+ */
+export declare function makeGateDecision(score: number): GateDecision;
+/**
+ * Calculate quality score from validation results
+ */
+export declare function calculateQualityScore(validationResults: Array<{
+    valid: boolean;
+    score: number;
+}>): number;
+//# sourceMappingURL=quality-gate.d.ts.map

package/dist/agents/quality-gate.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"quality-gate.d.ts","sourceRoot":"","sources":["../../src/agents/quality-gate.ts"],"names":[],"mappings":"AAAA;;;;;;GAMG;AAEH,OAAO,KAAK,EAAgB,YAAY,EAAE,MAAM,iBAAiB,CAAC;AAGlE;;GAEG;AACH,eAAO,MAAM,uBAAuB;;;;;;;;;;;;;;;;;;;CAmB1B,CAAC;AAEX;;GAEG;AACH,wBAAgB,uBAAuB,CACrC,IAAI,EAAE,MAAM,EACZ,EAAE,EAAE,MAAM,EACV,OAAO,EAAE;IACP,YAAY,CAAC,EAAE,MAAM,CAAC;IACtB,MAAM,CAAC,EAAE,MAAM,CAAC;IAChB,KAAK,CAAC,EAAE,MAAM,CAAC;CAChB,GACA;IAAE,KAAK,EAAE,OAAO,CAAC;IAAC,MAAM,EAAE,MAAM,EAAE,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CA4ErD;AAED;;GAEG;AACH,wBAAgB,gBAAgB,CAAC,KAAK,EAAE,MAAM,GAAG,YAAY,CAqB5D;AAED;;GAEG;AACH,wBAAgB,qBAAqB,CACnC,iBAAiB,EAAE,KAAK,CAAC;IAAE,KAAK,EAAE,OAAO,CAAC;IAAC,KAAK,EAAE,MAAM,CAAA;CAAE,CAAC,GAC1D,MAAM,CAKR"}

package/dist/agents/spec.d.ts

@@ -0,0 +1,4 @@
+import type { AgentConfig } from '../config/types';
+export declare const specAgentConfig: AgentConfig;
+export declare const specAgentPrompt = "You are SPEC - a structured specification agent following Hierarchical and Plan-and-Solve architectures with EARS/GEARS notation and phase-gated workflow.\n\nYOUR MISSION:\nTransform requirements into comprehensive, actionable specifications using EARS (Easy Approach to Requirements Syntax) notation, structured phase-gated workflow, and industry-best practices from Kiro and other leading AI coding tools.\n\nWORKFLOW (Phase-Gated, Hierarchical + Plan-and-Solve):\n\n### Phase 1: Requirements Generation\nGenerate requirements in EARS notation: WHEN [trigger] THE SYSTEM SHALL [response]\n1. **Business Analysis**: Understand user needs, business goals, constraints\n2. **User Stories**: Define user roles and their goals\n3. **EARS Requirements**: Write at least 3 requirements in EARS format\n   - Format: WHEN [trigger condition] THE SYSTEM SHALL [expected response]\n   - Each requirement must have acceptance criteria\n4. **Non-Functional Requirements**: Performance, Security, Usability\n\n**Phase Gate (requirements \u2192 design)**: Min 3 EARS requirements + acceptance criteria\n\n### Phase 2: Intelligence Gathering\nCall subagents in parallel for technical context:\n- **librarian**: Technical documentation and standards\n- **oracle**: Code analysis and architecture patterns\n- **architect**: System design and dependencies\n\n### Phase 3: Design Generation\nCreate comprehensive technical design:\n5. **Architecture Design**: Component diagram, data flow, patterns\n6. **Technology Stack**: Selected technologies with justification\n7. **Data Model**: Entity definitions, relationships\n8. **API Design**: Endpoints, request/response schemas\n9. **Error Handling**: Error codes, fault tolerance strategy\n10. **Security Design**: Auth, input validation, encryption\n\n**Phase Gate (design \u2192 tasks)**: Architecture diagram + technology stack\n\n### Phase 4: Task Generation\nCreate traceable implementation tasks:\n11. **Task Decomposition**: Break into phases (foundation \u2192 core \u2192 polish)\n12. **Dependency Mapping**: Define task dependencies and order\n13. **Requirement Traceability**: Link each task to EARS requirements (REQ-001)\n14. **Estimates**: Provide time estimates for each task\n\n**Phase Gate (tasks \u2192 implementation)**: Min 1 task + requirement traceability\n\n### Phase 5: Quality Review\n15. **Reviewer Subagent**: Validate spec completeness and consistency\n16. **Gate Validation**: Confirm all phase gates are met\n\nSPEC FORMATS:\n\n### EARS Notation (Requirements)\n```markdown\n## 4. Functional Requirements (EARS Format)\n\n| ID | Priority | Requirement |\n|----|----------|-------------|\n| REQ-001 | high | WHEN user submits valid credentials THE SYSTEM SHALL authenticate and create session |\n| REQ-002 | medium | WHEN authentication fails THE SYSTEM SHALL return error with retry guidance |\n| REQ-003 | high | WHEN session expires THE SYSTEM SHALL redirect to login with session timeout message |\n```\n\n### GEARS Extension\n```markdown\nWHEN [trigger] THE [component/service] SHALL [response]\n```\n\nOUTPUT FORMAT:\n\n# Specification Document: [Feature Name]\n> Spec Version: 1.0 | Phase: requirements | Workflow: requirements-first\n\n## \uD83D\uDCCB Requirements (EARS Format)\n\n### 1.1 Business Background\n[Project background and goals]\n\n### 1.2 User Stories\n- As a [role], I want [feature], so that [value]\n\n### 1.3 EARS Requirements\n| ID | Priority | Type | Trigger | Response | Acceptance Criteria |\n|----|----------|------|---------|----------|---------------------|\n| REQ-001 | high | WHEN | [trigger] | [response] | [criteria] |\n| REQ-002 | medium | WHEN | [trigger] | [response] | [criteria] |\n\n### 1.4 Acceptance Criteria\n- REQ-001: [Testable criteria]\n- REQ-002: [Testable criteria]\n\n### 1.5 Non-Functional Requirements\n- **Performance**: [Requirements]\n- **Security**: [Requirements]\n\n## \uD83C\uDFD7\uFE0F Technical Design\n\n### 2.1 Architecture\n```mermaid\ngraph TD\n    A[Client] --> B[API Gateway]\n    B --> C[Service Layer]\n    C --> D[Data Layer]\n```\n\n### 2.2 Technology Stack\n| Layer | Technology | Justification |\n|-------|-----------|---------------|\n| Frontend | [tech] | [reason] |\n| Backend | [tech] | [reason] |\n\n### 2.3 Data Model\n```typescript\ninterface Entity {\n  id: string;\n  // ...fields\n}\n```\n\n### 2.4 API Design\n| Method | Path | Description |\n|--------|------|-------------|\n| GET | /api/... | [description] |\n\n### 2.5 Error Handling\n[Error categories and handling strategy]\n\n## \uD83D\uDCC5 Implementation Tasks\n\n### 3.1 Phase 1: Foundation\n| Task ID | Description | REQ Link | Priority | Dependencies |\n|---------|-------------|----------|----------|-------------|\n| TASK-001 | [task] | REQ-001 | high | None |\n\n### 3.2 Phase 2: Core Features\n| Task ID | Description | REQ Link | Priority | Dependencies |\n|---------|-------------|----------|----------|-------------|\n\n### 3.3 Phase 3: Integration\n| Task ID | Description | REQ Link | Priority | Dependencies |\n|---------|-------------|----------|----------|-------------|\n\n### 3.4 Traceability Matrix\n| Requirement | Tasks | Coverage |\n|-------------|-------|----------|\n| REQ-001 | TASK-001 | full |\n\n## \u2705 Verification & Quality Gates\n\n### 4.1 Phase Gate Status\n- [x] Requirements: 3+ EARS items, acceptance criteria\n- [ ] Design: Architecture + tech stack\n- [ ] Tasks: 1+ tasks, requirement traceability\n\n### 4.2 Testing Strategy\n- Unit tests: >80% coverage\n- Integration tests: Critical paths\n- E2E tests: Core workflows\n\n---\n\n**Next Phase**: [design|tasks|implementation]\n**Phase Gate**: [pass/fail - resolution required]\n\n=== EARS/GEARS NOTATION RULES ===\n\n**EARS Format Requirements**:\n- Every functional requirement MUST use EARS notation\n- Format: WHEN [trigger] THE SYSTEM SHALL [response]\n- GEARS extension: WHEN [trigger] THE [component] SHALL [response]\n- Minimum 3 EARS-formatted requirements\n- Each requirement MUST have acceptance criteria\n\n**Phase Gate Rules**:\n- requirements \u2192 design: Min 3 EARS requirements + acceptance criteria\n- design \u2192 tasks: Architecture diagram + technology stack\n- tasks \u2192 implementation: Min 1 task + requirement traceability (REQ links)\n\n**Traceability Rules**:\n- Every task MUST link to at least one requirement ID\n- Every requirement SHOULD be covered by at least one task\n- Use REQ-001 format for requirement references in tasks\n\n=== ARCHITECTURE PATTERN RULES ===\n\n**Hierarchical Approach**:\n- Break work into phases (requirements \u2192 design \u2192 tasks)\n- Each phase dependent on previous phase completion\n- Subagents called in parallel where possible\n\n**Plan-and-Solve Principles**:\n- Complete planning before execution\n- Structured task decomposition with dependencies\n- Clear acceptance criteria for each task\n\n**Quality Rules**:\n- All EARS requirements validated for syntax\n- All phase gates checked before advancement\n- Traceability maintained throughout spec lifecycle";
+//# sourceMappingURL=spec.d.ts.map

package/dist/agents/spec.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"spec.d.ts","sourceRoot":"","sources":["../../src/agents/spec.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAC;AAEnD,eAAO,MAAM,eAAe,EAAE,WAkB7B,CAAC;AAEF,eAAO,MAAM,eAAe,u3NAoMwB,CAAC"}

package/dist/agents/vibe.d.ts

@@ -0,0 +1,4 @@
+import type { AgentConfig } from '../config/types';
+export declare const vibeAgentConfig: AgentConfig;
+export declare const vibeAgentPrompt = "You are VIBE - a rapid prototyping agent following the Reflector architecture pattern.\n\nYOUR MISSION:\nTransform ideas into executable prototypes through creative exploration and quick implementation with continuous self-improvement.\n\nWORKFLOW (Reflector Pattern):\n\n### Phase 1: Generate Initial Prototype\n1. **Intent Recognition**: Identify Vibe style (rapid_prototype/inspiration_capture/exploratory_coding)\n2. **Inspiration**: Call inspirer subagent to generate creative solutions\n3. **Context**: Call shared subagents (librarian+oracle) for technical context\n4. **Generate**: Create initial executable prototype code\n\n### Phase 2: Evaluate Quality\n5. **Self-Assessment**: Evaluate prototype against criteria:\n   - Code quality and readability\n   - Technical feasibility\n   - User experience\n   - Performance considerations\n6. **Validation**: Call explorer subagent to validate feasibility\n\n### Phase 3: Refine and Improve\n7. **Identify Gaps**: Detect areas for improvement\n8. **Apply Fixes**: Implement refinements\n9. **Iterate**: Return to Phase 1 if quality threshold not met\n\nVIBE STYLES:\n- **Rapid Prototype**: Quick implementation, loose style, suitable for iteration\n- **Inspiration Capture**: Idea capture, focus on quick realization\n- **Exploratory Coding**: Try and adjust, suitable for experimental projects\n\nOUTPUT FORMAT:\n# Vibe Prototype: [Project/Feature Name]\n\n## \uD83D\uDCA1 Inspiration Source\n[Core creative idea and technical approach]\n\n## \uD83D\uDD04 Reflector Assessment\n**Iteration**: [Number]\n**Quality Score**: [1-10]\n**Areas for Improvement**: [List]\n\n## \uD83D\uDE80 Rapid Prototype\n```[language]\n// Executable prototype code\n```\n\n## \uD83D\uDCCB Next Steps\n1. [Optimization direction 1]\n2. [Optimization direction 2]\n3. [Optimization direction 3]\n\n## \uD83D\uDD27 Technical Points\n- **Core Tech Stack**: [Technologies used]\n- **Key Dependencies**: [Required libraries/frameworks]\n- **Considerations**: [Issues to note]\n\n---\n**Tip**: Switch to Spec mode for complete specification document\n\n=== ARCHITECTURE PATTERN: REFLECTOR ===\n\n**Generate**: Create initial solution based on inspiration and context\n**Evaluate**: Assess quality against technical and user criteria\n**Improve**: Apply refinements and iterate until quality threshold met\n\n**Memory Components**:\n- Track iterations and improvements\n- Store quality metrics and assessment results\n- Maintain history of design decisions\n\n=== RULES ===\n- ALL paths from subagents\n- Be specific, never vague\n- Focus on executability\n- Embrace creativity\n- Apply self-reflection and improvement\n- Iterate until quality threshold is met";
+//# sourceMappingURL=vibe.d.ts.map

package/dist/agents/vibe.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"vibe.d.ts","sourceRoot":"","sources":["../../src/agents/vibe.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,WAAW,EAAE,MAAM,iBAAiB,CAAC;AAEnD,eAAO,MAAM,eAAe,EAAE,WAiB7B,CAAC;AAEF,eAAO,MAAM,eAAe,ooFA6Ea,CAAC"}

package/dist/config/agent-configs.d.ts

@@ -0,0 +1,5 @@
+import type { AgentConfig, ModelConfig } from './types';
+export declare const primaryAgentConfigs: Record<string, Partial<AgentConfig>>;
+export declare const subagentConfigs: Record<string, Partial<AgentConfig>>;
+export declare const modelRecommendations: Record<string, ModelConfig>;
+//# sourceMappingURL=agent-configs.d.ts.map

package/dist/config/agent-configs.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"agent-configs.d.ts","sourceRoot":"","sources":["../../src/config/agent-configs.ts"],"names":[],"mappings":"AAAA,OAAO,KAAK,EAAE,WAAW,EAAE,WAAW,EAAE,MAAM,SAAS,CAAC;AAIxD,eAAO,MAAM,mBAAmB,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,WAAW,CAAC,CA8CpE,CAAC;AAGF,eAAO,MAAM,eAAe,EAAE,MAAM,CAAC,MAAM,EAAE,OAAO,CAAC,WAAW,CAAC,CAqFhE,CAAC;AAGF,eAAO,MAAM,oBAAoB,EAAE,MAAM,CAAC,MAAM,EAAE,WAAW,CAe5D,CAAC"}

package/dist/config/constants.d.ts

@@ -0,0 +1,136 @@
+export declare const AGENT_MODES: readonly ["vibe", "spec", "ask"];
+export declare const PRIMARY_AGENTS: readonly ["vibe", "spec", "ask"];
+export declare const SUBAGENTS: readonly ["librarian", "oracle", "architect", "inspirer", "explorer", "planner", "reviewer", "quality-analyzer", "quality-reviewer", "quality-metrics", "quality-reporter"];
+export declare const TODO_ENFORCER_AGENTS_MANDATORY: string[];
+export declare const TODO_ENFORCER_AGENTS_ADVISORY: string[];
+export declare const TODO_MARKER_START = "<!-- TODOS-START -->";
+export declare const TODO_MARKER_END = "<!-- TODOS-END -->";
+export declare const QUALITY_PROMOTE_THRESHOLD = 8;
+export declare const QUALITY_HOLD_THRESHOLD = 6;
+export declare const QUALITY_ROLLBACK_THRESHOLD = 3;
+export declare const QUALITY_BASE_SCORE = 5;
+export declare const NEGATIVE_PENALTY_CRITICAL = 3;
+export declare const NEGATIVE_PENALTY_MAJOR = 2;
+export declare const NEGATIVE_PENALTY_MINOR = 1;
+export declare const STORE_MAX_AGE_MS: number;
+export declare const STORE_CLEANUP_INTERVAL_MS: number;
+export declare const PLUGIN_NAME = "opencode-vibe-spec-plugin";
+export declare const PLUGIN_VERSION = "1.0.4";
+export declare const PLUGIN_LOG_SERVICE = "vibe-spec";
+export declare const DEFAULT_MODEL = "zhipu/glm-4-flash";
+export declare const DEFAULT_PRIMARY_MODEL = "zhipu/glm-5.2";
+export declare const AGENT_COLORS: {
+    readonly vibe: "#FF6B35";
+    readonly spec: "#4A90E2";
+    readonly ask: "#9C27B0";
+};
+export declare const AGENT_STEPS: {
+    readonly vibe: 12;
+    readonly spec: 25;
+    readonly ask: 20;
+};
+export declare const VALID_MODEL_PATTERNS: readonly [RegExp, RegExp, RegExp, RegExp, RegExp, RegExp, RegExp, RegExp, RegExp, RegExp, RegExp, RegExp];
+export declare const SECURITY_PATTERNS: readonly [{
+    readonly pattern: RegExp;
+    readonly severity: "critical";
+    readonly cwe: "CWE-95";
+    readonly owasp: "A03:2021";
+    readonly message: "eval() usage detected";
+    readonly recommendation: "Remove eval() and use safe alternatives";
+}, {
+    readonly pattern: RegExp;
+    readonly severity: "high";
+    readonly cwe: "CWE-79";
+    readonly owasp: "A03:2021";
+    readonly message: "innerHTML assignment detected";
+    readonly recommendation: "Use textContent or safe DOM manipulation";
+}, {
+    readonly pattern: RegExp;
+    readonly severity: "high";
+    readonly cwe: "CWE-79";
+    readonly owasp: "A03:2021";
+    readonly message: "document.write() usage detected";
+    readonly recommendation: "Use DOM manipulation methods instead";
+}, {
+    readonly pattern: RegExp;
+    readonly severity: "critical";
+    readonly cwe: "CWE-95";
+    readonly owasp: "A03:2021";
+    readonly message: "new Function() usage detected";
+    readonly recommendation: "Remove dynamic function creation";
+}, {
+    readonly pattern: RegExp;
+    readonly severity: "critical";
+    readonly cwe: "CWE-798";
+    readonly owasp: "A07:2021";
+    readonly message: "Hardcoded password detected";
+    readonly recommendation: "Use environment variables for secrets";
+}, {
+    readonly pattern: RegExp;
+    readonly severity: "critical";
+    readonly cwe: "CWE-798";
+    readonly owasp: "A07:2021";
+    readonly message: "Hardcoded API key detected";
+    readonly recommendation: "Use environment variables for secrets";
+}, {
+    readonly pattern: RegExp;
+    readonly severity: "critical";
+    readonly cwe: "CWE-798";
+    readonly owasp: "A07:2021";
+    readonly message: "Hardcoded secret detected";
+    readonly recommendation: "Use environment variables for secrets";
+}, {
+    readonly pattern: RegExp;
+    readonly severity: "critical";
+    readonly cwe: "CWE-89";
+    readonly owasp: "A03:2021";
+    readonly message: "SQL injection pattern detected";
+    readonly recommendation: "Use parameterized queries";
+}, {
+    readonly pattern: RegExp;
+    readonly severity: "critical";
+    readonly cwe: "CWE-78";
+    readonly owasp: "A03:2021";
+    readonly message: "Command injection pattern detected";
+    readonly recommendation: "Use safe command execution";
+}, {
+    readonly pattern: RegExp;
+    readonly severity: "high";
+    readonly cwe: "CWE-79";
+    readonly owasp: "A03:2021";
+    readonly message: "React dangerouslySetInnerHTML usage";
+    readonly recommendation: "Sanitize HTML content before rendering";
+}];
+export declare const COMPLEXITY_THRESHOLDS: {
+    readonly cyclomatic: {
+        readonly excellent: 5;
+        readonly good: 10;
+        readonly acceptable: 20;
+        readonly poor: 50;
+    };
+    readonly cognitive: {
+        readonly excellent: 5;
+        readonly good: 10;
+        readonly acceptable: 15;
+        readonly poor: 25;
+    };
+    readonly functionLength: {
+        readonly excellent: 20;
+        readonly good: 50;
+        readonly acceptable: 100;
+        readonly poor: 200;
+    };
+    readonly nestingDepth: {
+        readonly excellent: 2;
+        readonly good: 3;
+        readonly acceptable: 5;
+        readonly poor: 8;
+    };
+    readonly duplication: {
+        readonly excellent: 0;
+        readonly good: 5;
+        readonly acceptable: 10;
+        readonly poor: 20;
+    };
+};
+//# sourceMappingURL=constants.d.ts.map

package/dist/config/constants.d.ts.map

@@ -0,0 +1 @@
+{"version":3,"file":"constants.d.ts","sourceRoot":"","sources":["../../src/config/constants.ts"],"names":[],"mappings":"AACA,eAAO,MAAM,WAAW,kCAAmC,CAAC;AAC5D,eAAO,MAAM,cAAc,kCAAmC,CAAC;AAC/D,eAAO,MAAM,SAAS,6KAKZ,CAAC;AAGX,eAAO,MAAM,8BAA8B,UAAW,CAAC;AACvD,eAAO,MAAM,6BAA6B,UAAU,CAAC;AAGrD,eAAO,MAAM,iBAAiB,yBAAyB,CAAC;AACxD,eAAO,MAAM,eAAe,uBAAuB,CAAC;AAGpD,eAAO,MAAM,yBAAyB,IAAI,CAAC;AAC3C,eAAO,MAAM,sBAAsB,IAAI,CAAC;AACxC,eAAO,MAAM,0BAA0B,IAAI,CAAC;AAG5C,eAAO,MAAM,kBAAkB,IAAI,CAAC;AAGpC,eAAO,MAAM,yBAAyB,IAAI,CAAC;AAC3C,eAAO,MAAM,sBAAsB,IAAI,CAAC;AACxC,eAAO,MAAM,sBAAsB,IAAI,CAAC;AAGxC,eAAO,MAAM,gBAAgB,QAAsB,CAAC;AACpD,eAAO,MAAM,yBAAyB,QAAiB,CAAC;AAGxD,eAAO,MAAM,WAAW,8BAA8B,CAAC;AACvD,eAAO,MAAM,cAAc,UAAU,CAAC;AACtC,eAAO,MAAM,kBAAkB,cAAc,CAAC;AAG9C,eAAO,MAAM,aAAa,sBAAsB,CAAC;AACjD,eAAO,MAAM,qBAAqB,kBAAkB,CAAC;AAGrD,eAAO,MAAM,YAAY;;;;CAIf,CAAC;AAGX,eAAO,MAAM,WAAW;;;;CAId,CAAC;AAGX,eAAO,MAAM,oBAAoB,2GAavB,CAAC;AAGX,eAAO,MAAM,iBAAiB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;EAWpB,CAAC;AAGX,eAAO,MAAM,qBAAqB;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;CAMxB,CAAC"}