fix(scoring): resolve gene symbols from var; never let a sample filter no-op

fix(sc,ui): filter lowly-expressed genes after pseudobulk; sweep every output dir

feat(sc): add decoupler_pseudobulk_aggregate — the missing Path P step

fix(scoring): close the six PROGENy per-sample gaps from the GSE205154 eval

refactor: move writable-dir constants to core.paths, fix workflow->tools layer violation

refactor: tighten invariant-length zips to strict=True (B905)

style: apply ruff lint --fix + ruff format across the tree

feat(integrity): ADR-0010 on-load verify for GEO series-matrix loads

feat(loveless): Role-2 signature consumable — resolve private signature URLs

feat(loveless): ADR-0006 + agent-side sc serving scaffold

feat(integration): ComBat per-sample scoring tool decoupler_pool_cohorts (ADR-0001 item 10)

refactor(loader): centralize authenticated URL->local resolution

ADR-0001 Phase 2 Mode A: early integration (T7 + T8 + T9)

feat(concordance): same-cohort normalization concordance routine (decoupler_normalization_concordance)

fix(loaders): accept http(s)/ftp URLs in bulk dataset loaders

feat(sanity): ADR-0002 Phase 2 — auto-fire tissue-identity contamination + PROGENy/Hallmark parity

feat(sanity): ADR-0002 Layer-2 result-aware sanity layer (Phase 1 + S1)

ADR-0001 Phase 1, step 4: meta-analysis engine (Mode B late integration)

Dedup: extract subset_and_require_group shared by activity_stats + microarray

Mark planned survival-stub params as intentionally unused (vulture)

Lint cleanup: remove unused imports, empty f-strings, split import lines

HF wheel registry, pinned deps, apt probe, security scan, API-key fallback

Fix 4 GPL-annotation fallback edge cases + 2 integration test bugs

Document residual GPL annotation limitations (Known Gaps #15)

Add GEOparse fallback for GPL platforms with no .annot.gz (Tier 2)

Switch GPL platform annotation to NCBI .annot.gz instead of _family.soft.gz

Replace rpy2 limma bridge with Rscript subprocess — fixes all conversion errors

Fix run_limma: rename R globals from _lm_ to .lm_ (R names can't start with underscore)

Rewrite run_limma: build design matrix in R to fix non-conformable arrays

Broaden rpy2 converter fallback to catch TypeError in addition to AttributeError

Fix limma: wrap all numpy→R and R→pandas ops in localconverter

Fix limma: replace deprecated rpy2 .activate() API with localconverter

Add dataset_validate_manifest_against_data tool and tests

Refactor manifest/metadata to separate semantics from runtime counts

Fix decode_geo_numeric_codes to decode all code segments, not just last

Remove dead code and standardize MCP tool return schemas

Added seperate files, added tests

Add PDAC dataset adapter, workflow, semantic, and analysis layers