Generated from EXPANSION Review 2026-03-16
- What: Add
--cov-fail-under=80to Makefile test target and CI workflow. - Why: No automated coverage gate exists. New devices can ship with low coverage.
- Effort: S (1 hour)
- Priority: P0
- What: Adopt consistent BDD naming convention across all device tests. Enforce
test_given_<scenario>_when_<action>_then_<result>pattern. Update _template test to demonstrate. - Why: Tests are inconsistent — bruker-2p uses BDD-style docstrings, biotek-gen5 uses plain names. Project convention mandates BDD.
- Effort: S (2 hours)
- Priority: P0
- What: Change bruker-2p adapter.py
password: str = "0000"topassword: str = os.getenv("PRAIRIE_PASSWORD", "0000"). Document why default exists. - Why: Sets bad pattern for future devices. Demo fallback is fine but should be explicit.
- Effort: S (1 hour)
- Priority: P0
- What: Create
devices/bruker-topspin/user/with MEMORY.md, system_config.md, calibration_log.md, protocols/, findings/. Mirror structure from bruker-2p and biotek-gen5. - Why: Two-layer RAG (docs/ + user/) is the standard template. bruker-topspin is missing user/.
- Effort: S (1 hour)
- Priority: P0
- What: Add
tests/test_cross_repo_compat.pythat verifies device-use can load adapters and labclaw can instantiate drivers from skill.yaml. - Why: Breaking changes to BaseDriver/BaseAdapter could slip through. No automated cross-repo verification exists.
- Effort: M (4 hours)
- Priority: P0
- What: Add
src/device_skills/validator.py—validate_device(device_dir)checks skill.yaml capabilities match actual adapter implementation. Integrate into loader. - Why: Metadata (skill.yaml) can drift from reality. Automated validation catches mismatches.
- Effort: M (3 hours)
- Priority: P1
- What: Add
logging.getLogger(__name__)to all modules. Standardize log levels. Add audit trail for adapter lifecycle. - Why: No observability for device operations. Required before production lab deployment.
- Effort: M (4 hours)
- Priority: P1
- What: Create
devices/micro-manager/with full driver/adapter/processor/brain implementation. Simulation-first imaging skill. - Why: Highest ROI expansion target per ROADMAP.md.
- Effort: XL (50 hours)
- Priority: P2