javiertarazon Polyglot Testing Pipeline
Autonomous 6-agent pipeline for comprehensive test generation across any programming language. Based on the Polyglot Test Generator pattern from github/awesome-copilot.
Supported Languages
Python · TypeScript · JavaScript · Go · Rust · Java · C# · Ruby · PHP · Swift · Kotlin
Pipeline Architecture
Polyglot Test Generator (Orchestrator)
│
├─► 1. Researcher — Analyzes existing code and test patterns
├─► 2. Planner — Designs test strategy and coverage goals
├─► 3. Implementer — Writes the actual test code
├─► 4. Tester — Runs tests and captures failures
├─► 5. Fixer — Diagnoses and fixes failing tests
└─► 6. Linter — Enforces style and test quality standards
Stage Descriptions
Stage 1 — Researcher
Input: Source file(s) to test
Output: Analysis report
Tasks:
- Identify functions, classes, methods to test
- Detect existing test patterns/frameworks in the project
- Find edge cases from type signatures and docstrings
- Map dependencies that need mocking
Detected framework: pytest (Python)
Test coverage: 23% (needs: 80%+)
Functions to test: 12 (5 pure, 7 with side effects)
Mocks needed: database.users, external_api.fetch
Stage 2 — Planner
Input: Researcher report
Output: Test plan with priorities
test_plan:
- id: "test_user_create"
priority: high
type: unit
inputs: [valid_data, invalid_email, duplicate_username]
expected: [201, 422, 409]
mocks: [db.users.insert]
- id: "test_auth_flow"
priority: critical
type: integration
steps: [register → login → access_protected → logout]
Stage 3 — Implementer
Input: Test plan
Output: Complete test file(s)
Writes tests following project conventions:
- Uses correct assertion style for the framework
- Implements fixtures/factories for test data
- Applies proper mocking patterns
- Groups tests logically (describe/context blocks)
Stage 4 — Tester
Input: Generated tests
Output: Test run results + failure report
- Executes the test suite
- Captures stdout/stderr
- Identifies flaky vs consistently failing tests
- Categorizes failures: assertion error, import error, timeout, etc.
Stage 5 — Fixer
Input: Failure report
Output: Fixed test file(s)
Fix strategies by error type:
| Error Type | Fix Strategy |
|---|---|
| ImportError | Add missing imports, check module path |
| AssertionError | Correct expected values from actual behavior |
| MockError | Fix mock setup, verify call signatures |
| TimeoutError | Add async handling, increase timeout |
| FixtureError | Fix fixture scope, add missing yields |
Stage 6 — Linter
Input: Fixed tests
Output: Final polished test file(s)
Enforces:
- Test naming conventions (
test_<action>_<condition>_<expected>) - No duplicate test logic
- Each test has exactly one assertion (where possible)
- All tests have docstrings
- Coverage target met (default 80%)
Framework Detection Rules
| Language | Frameworks (in priority order) |
|---|---|
| Python | pytest → unittest → nose2 |
| TypeScript/JS | Jest → Vitest → Mocha → Jasmine |
| Go | testing (std) → testify → ginkgo |
| Rust | cargo test → rstest |
| Java | JUnit 5 → TestNG |
| C# | xUnit → NUnit → MSTest |
| Ruby | RSpec → minitest |
Usage
Invoke full pipeline:
@workspace Run the polyglot testing pipeline on src/services/user.service.ts
Target coverage: 85%
Framework: Jest (TypeScript)
Invoke specific stage:
@workspace Run only Stage 3 (Implementer) using the attached test plan.
Add tests to existing suite:
@workspace The test file tests/test_auth.py exists.
Add coverage for the new password reset flow in src/auth/reset.py
Output Format
📦 Polyglot Test Pipeline — Complete
Language: Python | Framework: pytest
Source: src/services/user_service.py
📋 Plan: 8 test functions across 3 test classes
✅ Tests written: tests/test_user_service.py
🧪 Run results: 8 passed, 0 failed
🎯 Coverage: 87% (target: 80%) ✅
🔍 Lint: 0 warnings
Integration
- Use with
tdd-full-cyclewhen building new features - Use with
agent-orchestrationas the testing sub-agent - Use with
context-multi-filefor integration tests spanning multiple files