test-impact-mapper

name	test-impact-mapper
description	Map changed files to the smallest sufficient verification set. Use when: deciding which tests to run after a code change, explaining why a test scope is enough, or choosing between unit tests, integration tests, and manual checks.

Test-Impact Mapper

When to use this skill

• After any code change, to decide what to test.
• When the user asks "what tests should I run?" or "is this safe?"
• When reviewing a PR to assess test coverage of the diff.
• To justify skipping tests that cannot be affected by a change.

Test infrastructure

Runner	Command	Working dir	Env
CI (tests.yml)	pytest tests/ -v --tb=short --timeout=60	./services/chatbot	TESTING=True, MONGODB_ENABLED=False
CI (ci-cd.yml)	Same + lint job first	./services/chatbot	Same
Local	cd services/chatbot && pytest tests/ -v	services/chatbot	activate venv-core
RAG subsystem	cd rag && pytest tests/	rag	Python 3.11+
Security	bandit + pip-audit (CI only)	services/, app/src/	—

File-to-test map

Use this table to find the minimum test set for a changed file.

Changed file pattern	Tests to run	Reason
routes/stream.py	test_stream_complete_contract, test_stream_metrics_contract, test_app, test_api_integration	SSE contract and endpoint behavior
routes/main.py	test_app	Root, /chat, /clear, /history endpoints
routes/conversations.py	test_integration, test_repositories	Conversation CRUD + DB layer
routes/mcp.py	Manual: call MCP proxy routes	No dedicated unit test; verify via inspector
routes/image_gen.py	test_image_orchestration, test_endpoint_orchestrator_integration	Image generation pipeline
routes/memory.py	test_integration	Memory persistence through DB
routes/skills.py	test_skills_api	Skill management endpoints
core/chatbot.py	test_api_integration, test_integration, test_llm_clients, test_multi_turn_followup	Model routing and tool dispatch
core/chatbot_v2.py	Same as core/chatbot.py	V2 agent
core/tools.py	test_api_integration + manual search-tool check	Tool functions are mocked in CI; verify cascade manually
core/config.py	test_app (smoke)	Config is read at import; any test that imports the app exercises it
core/thinking_generator.py	test_agentic_* (all 9 files)	Thinking modes feed the agentic pipeline
core/streaming.py	test_stream_complete_contract, test_stream_metrics_contract	SSE helpers
core/extensions.py	test_app	Flask extensions shared across blueprints
core/db_helpers.py	test_integration, test_performance, test_repositories	Database access layer
core/error_handler.py	test_app	Error middleware
core/agentic/**	test_agentic_* (all 9 files)	Multi-agent orchestration
src/rag/**	test_ingest, test_retrieval, test_rag_*	RAG pipeline
src/audio_transcription.py	Manual only	No dedicated test
src/ocr_integration.py	Manual only	No dedicated test
src/video_generation.py	test_video_aspect_ratio_contract	Video parameter contract
templates/index.html	test_ui_smoke_contracts, test_app (root endpoint)	UI contract
static/**	test_ui_smoke_contracts	UI asset changes
services/shared_env.py	test_app (smoke) + check all services start	Env loader is process-wide
services/mcp-server/server.py	Inspector test: npx @modelcontextprotocol/inspector python server.py	MCP has no pytest suite
services/mcp-server/tools/**	Same inspector test	Utility functions not auto-registered
.env* / app/config/**	test_app (smoke)	Config read at startup
.github/workflows/**	No code test needed	Review YAML diff only
README.md / docs only	No test needed	Docs-only changes never need tests

Decision procedure

1. List changed files in the diff.
2. Look up each file in the map above.
3. Union the test sets — that is the minimum scope.
4. If any file has "Manual only", note it separately.
5. If the change is docs/config/workflow only, state "no tests needed" and explain why.
6. State the rationale: which tests cover which changed files, and why the remaining tests can be skipped.

Scope sufficiency rules

• Unit tests are enough when the change is limited to a single module and the file-to-test map covers it.
• Integration tests are needed when the change crosses module boundaries (e.g., route → chatbot → tool → config).
• Manual verification is needed when the map says "Manual only" or when the change affects MCP, audio, OCR, or live API integrations.
• Full suite (pytest tests/ -v) is warranted when: conftest.py changes, core/config.py changes, extensions.py changes, or more than 3 modules are affected.
• RAG tests are separate and only needed when src/rag/** changes.

CI workflow awareness

Workflow	Triggers on	What it validates
tests.yml	Push/PR to master, develop, feat/, refactor/	pytest tests/ -v in chatbot
ci-cd.yml	Same triggers	lint (compileall + flake8) → test-chatbot → test-root → security → docker
security-scan.yml	Push/PR to master, develop + weekly cron	bandit + pip-audit
codeql-analysis.yml	Push to master, develop + weekly cron	CodeQL security-and-quality
dependency-review.yml	PR to master, develop (requirements paths)	License + vulnerability check
rag-eval.yml	PR to main, feat/adv_RAG (app/rag/** paths)	RAG evaluation metrics

If CI will run the same tests automatically, say so. The user may still want to run locally first for faster feedback.

Mandatory checklist

After identifying the test scope:

• Every changed file is accounted for in the file-to-test map.
• The test set is the union of all matched tests — nothing dropped without justification.
• If a file maps to "Manual only," this is called out explicitly.
• The rationale explains why untouched tests can be skipped.
• If the change affects config, env, or extensions, the full suite is recommended.
• The response includes the exact command to run (e.g., pytest tests/test_stream_complete_contract.py tests/test_app.py -v).