---

name: tc-review description: Review QA test case documents (Markdown) against source code to find coverage gaps. Use when reviewing test cases, analyzing test coverage, suggesting missing scenarios, or learning codebase to extract testable scenarios. Supports project-based learning from multiple GitHub repos with dual memory (local + Dify RAG).

Test Case Review

Review test case documents against source code knowledge. Learn codebases, identify coverage gaps, suggest missing test cases.

Quick Start

# 1. Create project and add repos
python scripts/setup_project.py myproject https://github.com/org/repo.git

# 2. Review test cases
python scripts/review.py tests/checkout.md --project myproject

Decision Tree

User task → What do they need?
    │
    ├─ Review test cases → Is project set up?
    │   ├─ No → Run: python scripts/setup_project.py <name> <repo-url>
    │   └─ Yes → Run: python scripts/review.py <testcase.md> --project <name>
    │
    ├─ Learn new codebase → Run: python scripts/learn.py <project>
    │
    ├─ Get suggestions → Run: python scripts/suggest.py <project>
    │
    └─ Search knowledge → Run: python scripts/search.py <project> "<query>"

Available Scripts

Always run with --help first to see usage. Scripts handle complex workflows reliably.

Script	Purpose
scripts/setup_project.py	Create project + add repo + learn (all-in-one)
scripts/review.py	Review test case file against knowledge
scripts/learn.py	Learn/re-learn source code from repos
scripts/suggest.py	Generate test case suggestions
scripts/search.py	Search project knowledge
scripts/batch_review.py	Review multiple files at once

Test Case Format

Test cases must be Markdown. See references/testcase-format.md for complete guide.

# Test Suite: Payment Flow

## TC-001: Successful Payment
**Preconditions:**
- User logged in
- Cart has items

**Steps:**
1. Navigate to checkout
2. Enter valid card
3. Click Pay

**Expected Result:**
Payment success, order created

Example Workflows

Review Single File

python scripts/review.py tests/login.md --project auth-service

Output:

📊 Score: 65/100  Coverage: 65%

✅ Covered: Successful login, Invalid password
⚠️ Missing: Account locked, Rate limiting, Session timeout
💡 Suggested: TC-005 Account Lockout, TC-006 Rate Limit

Batch Review

python scripts/batch_review.py "tests/**/*.md" --project myapp --min-score 70

Quick Review (No Project Setup)

python scripts/setup_project.py temp-project https://github.com/org/repo.git
python scripts/review.py tests/feature.md --project temp-project

Environment Variables

Variable	Required	Description
ANTHROPIC_API_KEY	Yes*	For Claude LLM
OPENAI_API_KEY	Yes*	For OpenAI/Codex
LLM_PROVIDER	No	anthropic or openai (default: anthropic)
LLM_MODEL	No	Model name
DIFY_API_KEY	No	Enable Dify RAG
DIFY_DATASET_ID	No	Dify dataset ID

*One of ANTHROPIC_API_KEY or OPENAI_API_KEY required.

Reference Files

• testcase-format.md - Complete test case writing guide
• scenarios.md - Test scenario types and examples
• troubleshooting.md - Common issues and solutions