Test Suite Analysis

Test Suite Analysis Methodology

Extract behavioral intelligence from test suites. Tests are executable specifications -- they encode what the system MUST do in a form that can be verified. A passing test is a confirmed behavioral contract.

When to Use This Mode

Test suite analysis activates when:

• The target repository contains test files
• The discovery inventory identifies test files in the project
• Other modes discover test directories during analysis

This mode runs independently of all other intelligence sources. All output is RAW (test code references internal implementation details).

Why Tests Are High-Value Intelligence

Tests are the only source type that is simultaneously:

• Behavioral -- they describe what the system does, not how it's built
• Executable -- they can be run to confirm the behavior still holds
• Specific -- they provide exact inputs, expected outputs, and edge cases
• Maintained -- failing tests get fixed, so they track current behavior

A single end-to-end test is worth more than a page of documentation because the test is verified by CI on every commit.

Framework Detection

Identify the test framework(s) in use before analyzing test code. Different frameworks use different assertion styles, test organization, and execution models.

Detection Strategy

# Check for test configuration files
ls -la jest.config.* playwright.config.* cypress.config.* .rspec pytest.ini 2>/dev/null

# Check pyproject.toml for pytest config
grep -l '\[tool\.pytest' pyproject.toml 2>/dev/null

# Find test directories
find . -maxdepth 3 -type d \( -name "__tests__" -o -name "test" -o -name "tests" -o -name "spec" -o -name "cypress" \) 2>/dev/null

# Find test files by naming convention
find . -maxdepth 4 -type f \( -name "*.test.*" -o -name "*.spec.*" -o -name "test_*" -o -name "*_test.*" \) 2>/dev/null | head -50

# Count test files per pattern
echo "Jest/Mocha-style:" && find . -name "*.test.*" -o -name "*.spec.*" 2>/dev/null | wc -l
echo "Python-style:" && find . -name "test_*.py" -o -name "*_test.py" 2>/dev/null | wc -l
echo "Go-style:" && find . -name "*_test.go" 2>/dev/null | wc -l
echo "JUnit-style:" && find . -path "*/src/test/*" -name "*.java" 2>/dev/null | wc -l

Write detection results to workspace/raw/test-evidence/test-inventory.md.

Strategy 1: Read Test Code

Read test files directly and extract behavioral claims. This strategy always works -- it requires no working environment, no dependencies, and no execution.

1.1 Test File Inventory

# Build complete inventory of test files with metadata
find . -type f \( -name "*.test.*" -o -name "*.spec.*" -o -name "test_*" -o -name "*_test.*" -o -name "*_test.go" \) 2>/dev/null | while read f; do
  lines=$(wc -l < "$f")
  echo "$lines $f"
done | sort -rn

1.2 Assertion Extraction

For each test file, extract the assertions -- these are the behavioral contracts:

# Jest/Mocha assertions
grep -n "expect\|assert\|should\|toBe\|toEqual\|toContain\|toThrow\|toHaveBeenCalled" "$TEST_FILE"

# pytest assertions
grep -n "assert \|assert_\|assertEqual\|assertRaises\|pytest.raises" "$TEST_FILE"

# Go test assertions
grep -n "t\.Error\|t\.Fatal\|t\.Log\|assert\.\|require\." "$TEST_FILE"

# RSpec assertions
grep -n "expect\|should\|is_expected\|eq(\|include(\|raise_error" "$TEST_FILE"

1.3 Given/When/Then Extraction

Transform test code into behavioral claims using Given/When/Then structure:

For each test case (it(, test(, func Test*, def test_*), extract:

• Given (setup/preconditions): fixture creation, mock configuration, state initialization
• When (action): the function call, API request, or user action being tested
• Then (assertions): the expected outcomes encoded in assertions

## Test: "should reject expired tokens"

**Given:** A token with expiry date in the past
**When:** The token is validated via `checkPermissions()`
**Then:**
- Returns false
- Sets error to "TOKEN_EXPIRED"
- Does not call the downstream service

**Source:** `auth.test.ts:45-62`
**Confidence:** confirmed (test assertion is an explicit behavioral contract)

1.4 E2E vs Unit Value Classification

Not all tests carry equal behavioral intelligence value:

Focus extraction effort on e2e and integration tests first. Unit tests fill in details after the behavioral surface is mapped.

1.5 Edge Case Mining

Tests are the richest source of edge case documentation. Look for:

# Boundary value tests
grep -n "boundary\|limit\|max\|min\|overflow\|underflow\|zero\|empty\|null\|undefined" "$TEST_FILE" -i

# Error condition tests
grep -n "error\|fail\|reject\|throw\|invalid\|unauthorized\|forbidden\|timeout" "$TEST_FILE" -i

# Concurrency tests
grep -n "concurrent\|parallel\|race\|deadlock\|lock\|mutex\|async\|await" "$TEST_FILE" -i

# Special character and encoding tests
grep -n "unicode\|utf\|encoding\|escape\|special\|whitespace" "$TEST_FILE" -i

Write behavioral claims to workspace/raw/test-evidence/behavioral-claims.md. Write e2e flow documentation to workspace/raw/test-evidence/e2e-flows.md. Write edge case documentation to workspace/raw/test-evidence/edge-cases.md.

Strategy 2: Run Test Suite

Execute the test suite and observe its behavior. This strategy requires a working environment with all dependencies installed. It produces higher-confidence claims but has higher setup cost.

2.1 Prerequisites

Before attempting test execution:

• Verify the container has all dependencies installed
• Check for required environment variables or config files
• Look for test setup scripts (beforeAll, setUp, fixtures, factories)
• Identify tests that require external services (databases, APIs)

2.2 Execute with Maximum Verbosity

# Jest
npx jest --verbose --no-coverage 2>&1 | tee workspace/raw/test-evidence/run-output.txt

# pytest
python -m pytest -v --tb=long 2>&1 | tee workspace/raw/test-evidence/run-output.txt

# Go
go test -v ./... 2>&1 | tee workspace/raw/test-evidence/run-output.txt

# RSpec
bundle exec rspec --format documentation 2>&1 | tee workspace/raw/test-evidence/run-output.txt

2.3 Observe Runtime Behavior

During test execution, capture:

• Network calls -- tests that make HTTP requests reveal API contracts
• File I/O -- tests that read/write files reveal data format contracts
• Timing -- slow tests may indicate external dependency interaction
• Failures -- failed tests reveal behavioral regressions or environment-specific behavior
• Warnings -- deprecation warnings and lint output reveal upcoming behavioral changes

# Capture network activity during tests (if strace/dtrace available)
strace -e trace=network -f npx jest 2> workspace/raw/test-evidence/network-trace.txt

# Capture file I/O during tests
strace -e trace=file -f npx jest 2> workspace/raw/test-evidence/file-trace.txt