test: agent skills infrastructure and marker taxonomy audit (#727, #728)#742

Open

planetf1 wants to merge 42 commits intogenerative-computing:mainfrom

planetf1:test/audit-markers-727-728

Commits on Mar 28, 2026

test: add granularity marker taxonomy infrastructure (#727)
planetf1
committed
test: add audit-markers skill for test classification (#728)
planetf1
committed
chore: add CLAUDE.md and agent skills infrastructure
planetf1
committed
test: improve audit-markers skill quality and add resource predicates
planetf1
committed
chore: remove issue references from audit-markers skill
planetf1
committed
docs: align MARKERS_GUIDE.md with predicate factory pattern
planetf1
committed
fix: validate_skill.py schema mismatch and brittle YAML parsing
planetf1
committed
fix: migrate deprecated llm markers to e2e, add backend registry, update audit-markers skill
planetf1
committed
feat: add estimate-vram skill and fix MPS VRAM detection
planetf1
committed
refactor: fold estimate-vram into audit-markers skill
planetf1
committed
docs: drop isolation refs and fix RAM guidance in markers docs
planetf1
committed
docs: add legacy marker guidance for example files in audit-markers skill
planetf1
committed
refactor: remove require_ollama() predicate — redundant with backend marker
planetf1
committed
refactor: replace requires_heavy_ram gate with huggingface backend marker in examples conftest
planetf1
committed
refactor: replace ad-hoc bedrock skipif with require_api_key predicate
planetf1
committed
refactor: migrate legacy resource markers to predicates
planetf1
committed
test: skip collection gracefully when optional backend deps are missing
planetf1
committed
test: refine integration marker definition and apply audit fixes
planetf1
committed
test: add importorskip guards and optional-dep skip logic for examples
planetf1
committed
fix: convert example import errors to skips; add cpex importorskip guards
planetf1
committed
test: skip OTel-dependent tests when opentelemetry not installed
planetf1
committed
fix: use conservative heuristic for Apple Silicon GPU memory detection
planetf1
committed
test: add training memory signals to audit-markers skill; bump alora VRAM gate
planetf1
committed
fix: cache system capabilities result in examples conftest
planetf1
committed
fix: cache get_system_capabilities() result in test/conftest.py
planetf1
committed
fix: flush MPS memory pool in intrinsic test fixture teardown
planetf1
committed
fix: load LocalHFBackend model in config dtype to prevent float32 upcasting
planetf1
committed
test: remove --isolate-heavy process isolation and bump intrinsic VRAM gates
planetf1
committed
test: migrate legacy markers in test_intrinsics_formatters.py
planetf1
committed
test: add integration marker to test_dependency_isolation.py
planetf1
committed
docs: document OLLAMA_KEEP_ALIVE=1m as memory optimisation for unordered test runs
planetf1
committed
fix: suppress mypy name-defined for torch.Tensor after importorskip change
planetf1
committed
fix: ruff format huggingface.py from_pretrained args
planetf1
committed
fix: ruff format test_watsonx.py and test_huggingface_tools.py
planetf1
committed
refactor: remove requires_gpu, requires_heavy_ram, requires_gpu_isolation markers and handlers
planetf1
committed
refactor: remove --ignore-*-check override flags from conftest
planetf1
committed
refactor: remove requires_api_key marker; fix api backend group to match watsonx+bedrock markers
planetf1
committed
fix: address review

ajbozarth
authored and
planetf1
committed
test: mark test_image_block_in_instruction as qualitative
planetf1
committed
chore: commit .claude/settings.json with skillLocations for skill discovery
planetf1
committed
docs: broaden audit-markers skill description to cover diagnostic use cases
planetf1
committed
docs: add diagnostic mode to audit-markers skill for troubleshooting skip/resource issues
planetf1
committed