devflow-prompts 1.0.0

package/flows/devflow.4-rev-test-plan.prompt.md

@@ -0,0 +1,124 @@
+---
+description: Review test plan for coverage and edge cases
+---
+
+# /devflow.4-rev-test-plan
+
+## Description
+Review test plan to check coverage, traceability to tech doc, and missing edge cases.
+
+## Input Requirements
+
+**Input Type: File Paths (Required)**
+- Provide paths to your design file and test plan file
+- **At least one filename MUST contain task-id** (e.g., `FEAT-001-design.md`)
+  - Task-id is needed to name output files
+  - Can be anywhere in filename: `FEAT-001`, `TASK-123`, `BUG-456`
+- **Files can be located anywhere** (not limited to `/devflow-prompts`)
+- Examples:
+  - `~/FEAT-001-design.md ~/FEAT-001-test.md` ✅
+  - `/tmp/TASK-123-spec.md /tmp/tests.md` ✅
+  - `./design.md ./tests.md` ❌ (no task-id)
+- AI will read both files and extract task-id from filenames
+
+**Rule File**: `devflow.4-rev-test-plan.rule.md` from `/devflow-prompts/rules/` (Test plan review checklist)
+
+## Input Validation
+
+- **Design file**: Required, must exist
+- **Test plan file**: Required, must exist
+- **Coding rules**: `devflow.4-rev-test-plan.rule.md` is required
+
+**If validation fails**: Stop immediately and output error message.
+
+## Output Format
+1. **Coverage gaps** (map to tech doc sections/use cases)
+2. **Missing negative/edge cases**
+3. **Flaky test risks**
+4. **Revised version**: Full updated `{#task-id}-test.md`
+
+## Prompt
+
+```
+You are a Senior QA reviewer.
+Review the test plan for completeness, traceability to tech doc, and missing edge cases.
+
+INPUTS:
+1) Tech doc
+<<<
+{USER INPUT 1 - Can be a file path OR design document content}
+>>>
+
+2) Current {#task-id}-test.md
+<<<
+{USER INPUT 2 - Can be a file path OR test plan content}
+>>>
+
+INSTRUCTIONS:
+
+1. **Read Input Files**:
+   - The user will provide file paths.
+   - USE `view_file` tool to read them.
+   - IF inputs are NOT file paths, ask user to provide file paths.
+   - File paths can be anywhere in the filesystem, not limited to `/devflow-prompts`.
+
+2. **Extract Task ID** (for file path input):
+   - Extract task-id from filenames using pattern: `[A-Z]+-[0-9]+`
+   - Check both filenames for task-id
+   - Examples: `FEAT-001-design.md` → `FEAT-001`
+   - If no task-id found, ask user to provide it
+
+OUTPUT:
+1) Coverage gaps (map to tech doc sections/use cases)
+2) Missing negative/edge cases
+3) Flaky test risks (timing, external deps)
+4) Revised version:
+- Output ONLY the full updated markdown for {#task-id}-test.md (same template)
+
+RULES:
+- Keep headings unchanged.
+- Do not add tests for features not described in tech doc.
+```
+
+## Review Checklist
+
+### Coverage Check
+- [ ] Each UC in tech doc has tests?
+- [ ] Each API endpoint has tests?
+- [ ] Each validation rule has tests?
+- [ ] Each error case has tests?
+- [ ] Security requirements have tests?
+
+### Traceability
+- [ ] Test cases clearly map to requirements?
+- [ ] Test coverage matrix complete?
+
+### Negative/Edge Cases
+- [ ] Invalid inputs?
+- [ ] Missing resources (404)?
+- [ ] Permission denied (403)?
+- [ ] Concurrent access?
+- [ ] Timeout scenarios?
+- [ ] Resource limits?
+
+### Flaky Test Risks
+- [ ] Tests depend on timing?
+- [ ] Tests depend on external services?
+- [ ] Tests depend on random data?
+- [ ] Tests depend on execution order?
+
+### Test Data Quality
+- [ ] Fixtures realistic?
+- [ ] Data matrix covers all scenarios?
+- [ ] Setup/teardown clear?
+
+## Usage Example
+
+## Usage Example
+
+```bash
+/devflow.4-rev-test-plan /devflow-prompts/specs/FEAT-001/FEAT-001-design.md /devflow-prompts/specs/FEAT-001/FEAT-001-test.md
+```
+
+## Next Steps
+After review is done → run `/devflow.5-gen-test-code` to generate test code

package/flows/devflow.5-gen-test-code.prompt.md

@@ -0,0 +1,114 @@
+---
+description: Generate test code from test plan (git patch format)
+---
+
+# /devflow.5-gen-test-code
+
+## Description
+Generate test code from test plan, output as git unified diff (patch).
+
+## Input Requirements
+
+**Input Type: File Path (Required)**
+- Provide path to your test plan file
+- **Filename MUST contain task-id** (e.g., `FEAT-001-test.md`, `TASK-123.md`)
+  - Task-id is needed to name output test code patch
+  - Can be anywhere in filename: `FEAT-001`, `TASK-123`, `BUG-456`
+- **File can be located anywhere** (not limited to `/devflow-prompts`)
+- Examples:
+  - `~/Documents/FEAT-001-test.md` ✅
+  - `/tmp/TASK-123-spec.md` ✅
+  - `./my-BUG-456.md` ✅
+  - `~/test-plan.md` ❌ (no task-id)
+- AI will automatically read the file.
+
+**Task Mode (Optional, Default: NEW)**
+- `NEW`: Generating tests for new features.
+- `UPDATE`: Updating tests for modified code.
+- AI will extract task-id from filename
+
+**Rule File**: `devflow.5-gen-test-code.rule.md` from `/devflow-prompts/rules/` (Test code writing standards)
+
+## Input Validation
+
+- **Test plan file**: Required, must exist
+- **File name**: Must match pattern `{task-id}-test.md`
+- **Coding rules**: `devflow.5-gen-test-code.rule.md` is required
+
+**If validation fails**: Stop immediately and output error message.
+
+## Output Format
+
+**Git unified diff (patch)** - NOT markdown!
+
+**Save to**: `/devflow-prompts/specs/{#task-id}/{#task-id}-test.patch`
+
+## Prompt
+
+```
+You are a Senior SDET.
+Implement tests exactly as described in the test plan and following test coding rules.
+
+INPUT 1: {#task-id}-test.md
+<<<
+{USER INPUT HERE - Can be a file path OR test plan content}
+>>>
+
+INSTRUCTIONS:
+
+1. **Read Input File**:
+   - The user will provide a file path.
+   - USE `view_file` tool to read its content.
+   - IF input is NOT a file path, ask user to provide a file path.
+   - File paths can be anywhere in the filesystem, not limited to `/devflow-prompts`.
+
+2. **Extract Task ID** (for file path input):
+   - Extract task-id from filename using pattern: `[A-Z]+-[0-9]+`
+   - Examples: `FEAT-001-test.md` → `FEAT-001`
+   - If no task-id found, ask user to provide it
+
+3. **Check Task Mode**:
+   - **IF Mode is UPDATE**:
+     - Ensure existing tests are updated, NOT just new tests added.
+     - Add regression tests to verify core logic remains intact.
+   - **IF Mode is NEW**:
+     - Focus on coverage for the new feature.
+
+INPUT 2: Test coding rules (devflow.5-gen-test-code.rule.md)
+<<<
+{paste test coding rules}
+>>>
+
+OUTPUT FORMAT:
+- Provide ONLY a git-style patch (unified diff).
+- Include only changed/new files.
+- No explanations outside the diff.
+
+RULES:
+- Follow test coding rules (naming, structure, mocking).
+- Tests must be deterministic (no flaky timing).
+- Cover happy path + negative + edge cases listed in test plan.
+- Add minimal test helpers/fixtures if needed.
+- Ensure tests adhere to all standards in test coding rules file.
+- **For UPDATE Mode**:
+  - Do NOT remove existing tests unless logic is explicitly deprecated.
+  - If refactoring, tests must pass for both old (if compatible) and new logic.
+```
+
+## Rules/Guidelines
+1. ✅ **Follow test rules** - Naming, structure, mocking conventions
+2. ✅ **Deterministic** - No flaky timing, no random data
+3. ✅ **Complete coverage** - Happy + negative + edge cases
+4. ✅ **Minimal helpers** - Only add necessary helpers
+5. ✅ **Clean assertions** - Clear, specific assertions
+
+## Usage Example
+
+## Usage Example
+
+```bash
+/devflow.5-gen-test-code /devflow-prompts/specs/FEAT-001/FEAT-001-test.md
+```
+
+## Next Steps
+After applying patch → run `/devflow.5-rev-test-code` to review

package/flows/devflow.5-rev-test-code.prompt.md

@@ -0,0 +1,131 @@
+---
+description: Review test code patch for reliability and correctness
+---
+
+# /devflow.5-rev-test-code
+
+## Description
+Review test code patch to check correctness, reliability, and alignment with test plan.
+
+## Input Requirements
+
+**Input Type: File Paths (Required)**
+- Provide paths to your test plan file and test patch file
+- **At least one filename MUST contain task-id** (e.g., `FEAT-001-test.md`)
+  - Task-id is needed to name output files
+  - Can be anywhere in filename: `FEAT-001`, `TASK-123`, `BUG-456`
+- **Files can be located anywhere** (not limited to `/devflow-prompts`)
+- Examples:
+  - `~/FEAT-001-test.md ~/test.patch` ✅
+  - `/tmp/TASK-123-spec.md /tmp/test-code.diff` ✅
+  - `./tests.md ./patch.diff` ❌ (no task-id)
+- AI will read both files and extract task-id from filenames
+
+**Rule File**: `devflow.5-rev-test-code.rule.md` from `/devflow-prompts/rules/` (Test code review checklist)
+
+## Input Validation
+
+- **Test plan file**: Required, must exist
+- **Generated test patch**: Required, must be valid git diff format
+- **Coding rules**: `devflow.5-rev-test-code.rule.md` is required
+
+**If validation fails**: Stop immediately and output error message.
+
+## Output Format
+1. **Findings** (grouped)
+2. **Must-fix list**
+3. **Corrected patch**: New unified diff with fixes
+
+## Prompt
+
+```
+You are a Tech Lead + QA reviewer.
+Review the test patch for correctness, reliability, and alignment with the test plan.
+
+INPUTS:
+1) Test plan
+<<<
+{USER INPUT 1 - Can be a file path OR test plan content}
+>>>
+
+2) Generated test patch (diff)
+<<<
+{USER INPUT 2 - Can be a file path OR test patch content}
+>>>
+
+INSTRUCTIONS:
+
+1. **Read Input Files**:
+   - The user will provide file paths.
+   - USE `view_file` tool to read them.
+   - IF inputs are NOT file paths, ask user to provide file paths.
+   - File paths can be anywhere in the filesystem, not limited to `/devflow-prompts`.
+
+2. **Extract Task ID** (for file path input):
+   - Extract task-id from filenames using pattern: `[A-Z]+-[0-9]+`
+   - Check both filenames for task-id
+   - Examples: `FEAT-001-test.md` → `FEAT-001`
+   - If no task-id found, ask user to provide it
+
+OUTPUT:
+1) Findings (grouped):
+- Missing/incorrect assertions
+- Not matching test plan cases
+- Flaky risks (timers, random, external calls)
+- Over-mocking / under-mocking
+- Test data problems
+2) Must-fix list
+3) Produce a corrected patch:
+- Output ONLY a new unified diff patch that fixes issues (no commentary outside diff)
+
+RULES:
+- Prefer minimal change.
+- If a test is impossible with current design, propose the smallest seam (DI/interface) and include it in patch.
+```
+
+## Review Checklist
+
+### Alignment with Test Plan
+- [ ] All test cases implemented?
+- [ ] Test levels correct (unit/integration/E2E)?
+- [ ] Preconditions set up correctly?
+- [ ] Expected outcomes verified?
+
+### Assertions Quality
+- [ ] Specific assertions (not just assertTrue)?
+- [ ] All expected values checked?
+- [ ] Error messages checked?
+- [ ] Side effects verified?
+
+### Reliability
+- [ ] No timing dependencies?
+- [ ] No random data?
+- [ ] No external service calls (or properly mocked)?
+- [ ] No execution order dependencies?
+
+### Mocking
+- [ ] Appropriate mocking (not over/under)?
+- [ ] Mock expectations correct?
+- [ ] Integration tests use real dependencies?
+
+### Test Data
+- [ ] Realistic fixtures?
+- [ ] Setup/teardown clean?
+- [ ] No data leakage between tests?
+
+### Code Quality
+- [ ] Test naming clear?
+- [ ] Test structure AAA (Arrange-Act-Assert)?
+- [ ] No code duplication?
+- [ ] Helpers used appropriately?
+
+## Usage Example
+
+## Usage Example
+
+```bash
+/devflow.5-rev-test-code /devflow-prompts/specs/FEAT-001/FEAT-001-test.md /devflow-prompts/specs/FEAT-001/FEAT-001-test.patch
+```
+
+## Next Steps
+After reviewing test code → run `/devflow.6-analyze` to validate entire flow

package/flows/devflow.6-analyze.prompt.md

@@ -0,0 +1,240 @@
+---
+description: Validate requirement → design → code → test flow, find inconsistencies
+---
+
+# /devflow.6-analyze
+
+## Description
+Validate the entire flow from requirement → tech design → code → test, detect inconsistencies and gaps.
+
+## Input Requirements
+
+**Input Type: File Paths (Required)**
+- Provide paths to your requirement, design, and test plan files
+- **Task-id extraction**: AI will try to extract from filenames (optional)
+  - If found, will be used in analysis report
+  - If not found, analysis will proceed without task-id reference
+- **Files can be located anywhere** (not limited to `/devflow-prompts`)
+- Examples:
+  - `~/FEAT-001-req.md ~/FEAT-001-design.md ~/FEAT-001-test.md` ✅ (with task-id)
+  - `/tmp/requirement.txt /tmp/spec.md /tmp/tests.md` ✅ (without task-id)
+- Feature code (optional: file path)
+- Test code (optional: file path)
+- AI will read all files
+- **Task Mode**: `NEW` or `UPDATE` (Optional, defaults to `NEW`)
+
+## Input Validation
+
+- **Requirement file**: Required, must exist
+- **Design file**: Required, must exist
+- **Test plan file**: Required, must exist
+- **Feature code**: Optional
+- **Test code**: Optional
+
+**If validation fails**: Stop immediately and output error message.
+
+## Output Format
+1. **Alignment check** - Requirement ↔ Tech ↔ Code ↔ Test
+2. **Inconsistencies found**
+3. **Coverage gaps**
+4. **Recommendations** (prioritized)
+
+## Prompt
+
+```
+You are a Principal Engineer + QA Lead conducting a final review.
+Validate alignment and surface inconsistencies across the entire implementation flow.
+
+INPUTS:
+1) Requirement: {#task-id}-requirement.md
+<<<
+{USER INPUT 1 - Can be a file path OR requirement content}
+>>>
+
+2) Tech Design: {#task-id}-design.md
+<<<
+{USER INPUT 2 - Can be a file path OR design content}
+>>>
+
+3) Test Plan: {#task-id}-test.md
+<<<
+{USER INPUT 3 - Can be a file path OR test plan content}
+>>>
+
+4) Feature Code (optional - key snippets)
+<<<
+{USER INPUT 4 - Can be a file path OR code snippets}
+>>>
+
+5) Test Code (optional - key snippets)
+<<<
+{USER INPUT 5 - Can be a file path OR test code snippets}
+>>>
+
+6) Task Mode (Optional): NEW | UPDATE
+
+INSTRUCTIONS:
+
+1. **Read Input Files**:
+   - The user will provide file paths.
+   - USE `view_file` tool to read them.
+   - IF inputs are NOT file paths, ask user to provide file paths.
+   - File paths can be anywhere in the filesystem, not limited to `/devflow-prompts`.
+
+2. **Extract Task ID** (optional, for file path input):
+   - Try to extract task-id from filenames using pattern: `[A-Z]+-[0-9]+`
+   - If found, use it in analysis report title
+   - If not found, proceed without task-id reference
+
+TASKS:
+1) Alignment check:
+   - Does tech design implement all requirements?
+   - Does code implement all tech design specs?
+   - Do tests cover all requirements and tech design?
+   
+2) Find inconsistencies:
+   - Contradictions between docs
+   - Missing implementations
+   - Over-implementations (not in requirement)
+   
+3) Coverage gaps:
+   - Requirements not in tech design
+   - Tech design not in code
+   - Code not tested
+   - Edge cases not handled
+   - (**UPDATE Only**) Regression risks:
+     - Are existing tests updated?
+     - Is backward compatibility maintained?
+   
+4) Recommendations:
+   - P0: Critical issues (blocking)
+   - P1: Important issues (should fix)
+   - P2: Nice-to-have improvements
+
+OUTPUT:
+- Structured report with findings and recommendations
+- No code generation, this is read-only analysis
+```
+
+## Rules/Guidelines
+1. ✅ **Read-only** - No code generation, only analyze
+2. ✅ **Traceability** - Map requirements → design → code → test
+3. ✅ **Inconsistencies** - Detect contradictions
+4. ✅ **Coverage** - Check if coverage is sufficient
+5. ✅ **Prioritize** - P0/P1/P2 recommendations
+
+## Analysis Checklist
+
+### Requirement → Tech Design
+- [ ] Each FR has implementation in tech design?
+- [ ] Each NFR is addressed?
+- [ ] Open questions resolved?
+- [ ] Assumptions validated?
+
+### Tech Design → Code
+- [ ] Each API endpoint implemented?
+- [ ] Data model changes applied?
+- [ ] Business logic implemented correctly?
+- [ ] Security measures in place?
+- [ ] Observability hooks added?
+
+### Tech Design → Test Plan
+- [ ] Each UC has test cases?
+- [ ] Edge cases covered?
+- [ ] Security tests present?
+- [ ] Performance tests if needed?
+
+### Test Plan → Test Code
+- [ ] Each test case implemented?
+- [ ] Coverage sufficient?
+- [ ] Assertions correct?
+
+### Cross-cutting Concerns
+- [ ] Consistent naming across docs/code?
+- [ ] Error handling consistent?
+- [ ] Validation rules consistent?
+- [ ] Data types consistent?
+
+## Usage Example
+
+## Usage Example
+
+```bash
+# 1. Run with file paths (can be anywhere in filesystem)
+/devflow.6-analyze \
+  /path/to/FEAT-001-requirement.md \
+  /path/to/FEAT-001-design.md \
+  /path/to/FEAT-001-test.md
+
+# 2. AI reads all files and analyzes
+# 3. Receive comprehensive analysis report
+```
+
+## Example Output Structure
+
+```markdown
+# Analysis Report: FEAT-001
+
+## 1. Alignment Check
+
+### Requirement → Tech Design ✅
+- All FRs mapped to tech design
+- All NFRs addressed
+- 2 open questions resolved in tech design
+
+### Tech Design → Code ⚠️
+- Issue: FR-3 rate limiting not fully implemented
+- Missing: Observability metrics for login attempts
+
+### Tech Design → Test Plan ✅
+- All UCs have test cases
+- Edge cases covered
+
+### Test Plan → Test Code ⚠️
+- Issue: TC-5 (concurrent login) not implemented
+- Coverage: 85% (target: 90%)
+
+## 2. Inconsistencies Found
+
+### P0 (Critical)
+- None
+
+### P1 (Important)
+- Tech design specifies 15-minute lock, code uses 10 minutes
+- Requirement says "show remaining attempts", code doesn't implement
+
+### P2 (Minor)
+- Naming inconsistency: "locked_until" vs "lockUntil" in different files
+
+## 3. Coverage Gaps
+
+- FR-3 rate limiting: Partially implemented
+- TC-5 concurrent login test: Not implemented
+- Metrics for observability: Missing
+
+## 4. Recommendations
+
+### P0 (Must fix before deploy)
+- None
+
+### P1 (Should fix)
+1. Fix lock duration: 10 min → 15 min (align with requirement)
+2. Implement "remaining attempts" display
+3. Implement TC-5 concurrent login test
+4. Add metrics hooks
+
+### P2 (Nice to have)
+1. Standardize naming: use "locked_until" everywhere
+2. Add more detailed logs for debugging
+3. Consider adding integration test for full flow
+```
+
+## When to Run
+
+- ✅ **Before final PR** - Catch issues early
+- ✅ **After major changes** - Verify alignment still holds
+- ✅ **Before deploy** - Final sanity check
+- ✅ **During code review** - Help reviewers spot gaps
+
+## Next Steps
+After analyze → Fix issues by priority → Ready to deploy!

package/gen-chat.sh

@@ -0,0 +1,58 @@
+#!/bin/bash
+
+SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
+SCRIPTS_DIR="$SCRIPT_DIR/scripts"
+
+# Function to show usage
+show_usage() {
+    echo "Usage: $0 [ide]"
+    echo "Available IDEs: vscode, cursor, antigravity"
+    echo "If no argument is provided, the script attempts to auto-detect the IDE."
+}
+
+# 1. Start with provided argument (converted to lowercase)
+IDE=$(echo "$1" | tr '[:upper:]' '[:lower:]')
+
+# 2. If no argument, try auto-detection
+if [ -z "$IDE" ]; then
+    # Helper for detection (checks TERM_PROGRAM env var)
+    TP=$(echo "$TERM_PROGRAM" | tr '[:upper:]' '[:lower:]')
+    
+    if [[ "$TP" == *"vscode"* ]]; then
+        IDE="vscode"
+    elif [[ "$TP" == *"cursor"* ]]; then
+        IDE="cursor"
+    # Fallback/Custom checks for Antigravity could go here if there's a known env var
+    # For now, we rely on the argument for Antigravity unless explicitly known
+    fi
+fi
+
+# 3. If still empty, error out
+if [ -z "$IDE" ]; then
+    echo "Error: Could not detect IDE and no argument provided."
+    show_usage
+    exit 1
+fi
+
+echo "Detailed setup for IDE: $IDE"
+
+case "$IDE" in
+    vscode)
+        TARGET_SCRIPT="$SCRIPTS_DIR/setup-devflow-chat-vscode.sh"
+        ;;
+    antigravity)
+        TARGET_SCRIPT="$SCRIPTS_DIR/setup-devflow-chat-antigravity.sh"
+        ;;
+    *)
+        echo "Error: Unsupported IDE '$IDE'"
+        show_usage
+        exit 1
+        ;;
+esac
+
+if [ -f "$TARGET_SCRIPT" ]; then
+    bash "$TARGET_SCRIPT"
+else
+    echo "Error: Setup script not found at $TARGET_SCRIPT"
+    exit 1
+fi

package/index.js

@@ -0,0 +1,4 @@
+#!/usr/bin/env node
+
+console.log("DevFlow Prompts is a CLI tool. Please run 'npx devflow-prompts init' or see README.md for usage.");
+module.exports = {};

package/package.json

@@ -0,0 +1,25 @@
+{
+  "name": "devflow-prompts",
+  "version": "1.0.0",
+  "description": "DevFlow Prompts - Prompt Engineering & Workflow Automation",
+  "main": "index.js",
+  "bin": {
+    "devflow-prompts": "./bin/cli.js"
+  },
+  "scripts": {
+    "test": "echo \"Error: no test specified\" && exit 1"
+  },
+  "keywords": [
+    "devflow",
+    "prompt",
+    "workflow",
+    "automation"
+  ],
+  "author": "Phong Le",
+  "license": "ISC",
+  "dependencies": {
+    "fs-extra": "^11.2.0",
+    "chalk": "^4.1.2",
+    "inquirer": "^8.2.6"
+  }
+}

package/raws/changename.md

@@ -0,0 +1,7 @@
+# Yêu cầu thay đổi tên dự án và prefix
+
+Thực hiện các thay đổi sau cho hệ thống prompt:
+
+1.  **Đổi tên dự án**: Cập nhật tên thư mục hoặc tên gói thành `devflow-prompts`.
+2.  **Đổi Prefix Prompt**: Cập nhật tiền tố (prefix) cho các lệnh prompt từ tên cũ sang `devflow`.
+    - Ví dụ: đổi thành `devflow.1-gen-requirement`, `devflow.2-gen-design`, v.v.