QA Review — MC #9697 reviewing MC #8941

QA Review — MC #9697 (reviewing MC #8941)

Reviewer: John / pi-orchestrator (Proveo persona)
Reviewed at (UTC): 2026-07-28T21:27:40Z
Source task: #8941 — ACS Phase 1.3 — Python SDK alai-config-py (Pydantic v2)
Source status: done
Source priority: H
Source owner: codecraft
Review task: #9697 — QA review: #8941

Tool-verified source task state

node ~/system/tools/mc.js show 8941 reports:

• Status: done
• Priority: H
• Owner: codecraft
• DoD/agent output: Python SDK alai-config-py built: 40/40 tests pass, wheel+sdist built, smoke test passes, AppConfig.load() typed model works. Proveo validation pending.

node ~/system/tools/mc.js show 9697 reports the QA review task was blocked because autowork had 5 failed attempts and needed human review.

The MC tool emitted [alai/config] Mock mode active from /Users/makinja/system/tools/alai-config-mock.json; this review is therefore anchored in the local/mock MC record plus direct file/test verification below.

Scope reviewed

Canonical package path reviewed:

• /Users/makinja/business/ALAI-Holding-AS/internal/packages/alai-config-py

Files read/reviewed:

• pyproject.toml
• README.md
• Makefile
• src/alai_config/__init__.py
• src/alai_config/client.py
• src/alai_config/resolver.py
• src/alai_config/models/app_config.py
• src/alai_config/cache.py
• src/alai_config/mock.py
• src/alai_config/errors.py
• src/alai_config/aliases.py
• tests/conftest.py
• tests/test_cache.py
• tests/test_mock.py
• tests/test_resolver.py

Lines reviewed by wc -l: 1,582 total across the files above.

Acceptance criteria check

Task 1.3 acceptance from /Users/makinja/projects/alai-system/specs/alai-config-system-plan.md:

Acceptance item	Review result	Evidence
Pydantic v2 AppConfig model generated from schema	PASS with implementation note	pyproject.toml requires pydantic>=2.0; src/alai_config/models/app_config.py defines Pydantic v2 models and AppConfig.load(). Makefile codegen supports regeneration from ~/ALAI/internal/alai-config/schema/app-config.schema.json; current model file states it is manually authored when the schema file is unavailable.
Refuses to start on invalid schema	PASS	AppConfig.load() calls cls.model_validate(raw) and re-raises validation errors after logging. Smoke verified typed model load; tests cover invalid mock JSON and config init failures.
PyPI publish alai-config	NOT VERIFIED	Local build produced wheel and sdist. I found no current machine evidence that the package was published to PyPI or a private package index.

Commands executed and evidence

Check	Command	Result	Evidence file
Unit tests	python3.12 -m pytest -q	PASS — 40 passed	/Users/makinja/system/evidence/9697/pytest.log
Package build	python3.12 -m build	PASS — built alai_config-1.0.0.tar.gz and alai_config-1.0.0-py3-none-any.whl	/Users/makinja/system/evidence/9697/build.log
Typed smoke	ALAI_CONFIG_MOCK_PATH=tests/fixtures/mock.json python3.12 ... AppConfig.load()	PASS — initialized true, endpoint http://localhost:3001, VAT rate 0.20 is Decimal, LightRAG URL http://localhost:9621	/Users/makinja/system/evidence/9697/smoke.log
Syntax compile	python3.12 -m compileall -q src tests	PASS	/Users/makinja/system/evidence/9697/static-checks.log
Secret pattern scan	grep -RInE ... src tests README.md pyproject.toml	PASS — no matches in the static check log	/Users/makinja/system/evidence/9697/static-checks.log
QA-19 source gate	node ~/system/tools/qa-19.js check 8941	PENDING — score 5/15; historic /tmp GOTCHA/verify markers absent	/Users/makinja/system/evidence/9697/qa19-8941.log

Build artifact hashes:

• dist/alai_config-1.0.0-py3-none-any.whl — SHA-256 69ec5e62fcbd0c87241161eae08e2cf9a362356f7ef38a9e44e1f1f9f4f885cb
• dist/alai_config-1.0.0.tar.gz — SHA-256 f9f46b830a6c2bbbc02a2cd262ae262ba6792ff64090135061a89fc3887cc9d6

Findings

Blocking defects

None found in the reviewed source and test scope.

Should-fix / follow-up

1. Publish evidence missing for the PyPI publish alai-config acceptance item.
   Local package build is verified, but no durable evidence was found for publish to PyPI or an internal package index. If publish was intentionally deferred, the acceptance item should be updated or a release task should be created.

2. QA-19 historic gate remains pending for #8941.
   Current direct verification passes, but qa-19.js check 8941 still fails because historic /tmp/gotcha-task-8941.md, /tmp/verify-8941, and hop-build marker files are absent. This review writes durable evidence under /Users/makinja/system/evidence/9697/ to close the review task, but it does not mutate the historic source task gate state.

Optional hardening

• Add explicit tests for AppConfig.load() invalid typed values, for example invalid VAT decimal or invalid Market enum, to directly prove the fail-fast Pydantic path beyond mock-file parse errors.
• Add a release checklist documenting the exact package index target for proprietary alai-config publishing.

Code review verdict

Status: APPROVED_WITH_COMMENTS
Security: PASS
Tests: PASS — 40/40
Build: PASS — wheel and sdist produced
Source QA-19: PENDING due missing historic /tmp evidence markers
Overall: The Python SDK implementation is usable and verified locally for tests, build, syntax, and typed AppConfig.load() smoke. The only acceptance gap I could not verify is package publication.

BookStack

Published and verified via BookStack API:

• URL: https://docs.alai.no/books/testing-qa/page/qa-review-mc-9697-reviewing-mc-8941
• Page ID: 3304

Required artifacts written by this review

• /Users/makinja/system/evidence/9697/qa-review-8941.md
• /Users/makinja/system/evidence/9697/summary.json
• /Users/makinja/system/evidence/9697/verification.json
• /Users/makinja/system/evidence/9697/pytest.log
• /Users/makinja/system/evidence/9697/build.log
• /Users/makinja/system/evidence/9697/smoke.log
• /Users/makinja/system/evidence/9697/static-checks.log
• /Users/makinja/system/evidence/9697/qa19-8941.log