feat(fusion): map modality predictions to per-disease signals

feat(fusion): add clinical-test signal normalisers (MMSE/MoCA/UPDRS/gait/age)

feat(fusion): add disease/modality weight registry

fix(fusion): drop unused protected_namespaces; pin tighter bound tests

feat(fusion): add pydantic data contract for multi-modal fusion

fix(agents/orchestrator): log dropped out-of-stage tool calls (was silent)

docs(agents/tools): TODO race on shared parquet output for concurrent /agent/run

feat(frontend/mri): expose target_shape as 3 number inputs (was hardcoded 64³)

fix(mri/model): warn when label_names length != model output dim (was silent override)

Add project files

feat(agents): live OpenRouter integration test (slow) + GET /diag/agent

feat(frontend): Agent tab with decision-trace expander

feat(api): POST /agent/run endpoint (orchestrator + RAG, stub-injectable)

fix(agents/tools): parameterize processed_dir + translate HTTPException → ValueError

feat(agents): orchestrator loop (function-calling + tool trace + max-steps gate)

fix(rag/ingest): drop print() in main() — INFO log already covers it (AGENTS.md §3)

feat(agents): Tool dataclass + registry + 4 tool wrappers (3 pipelines + RAG)

feat(rag): RAGRetriever (load + search → chunks with scores)

feat(rag): ingest CLI (markdown/PDF → chunks → FAISS) + sample KB fixtures

fix(rag/store): copy vectors before in-place normalize_L2 (no caller mutation)

feat(rag): FAISS inner-product store with chunk metadata + roundtrip

feat(rag): fastembed wrapper (Embedder, bge-small-en-v1.5, 384-dim)

feat(rag): paragraph-aware chunker (chunk_text)

feat(llm): user-question-driven prompt (language match + intent split)

chore(diag): add /diag/openrouter endpoint + sidebar button

docs(llm): add probe-date + 429-policy notes; fix chain capability order

feat(llm): refresh free-tier chain with verified-live OpenRouter IDs

feat(llm): classify 401 as fatal+actionable, 400 as skip-this-model

feat(llm): OpenRouter free-tier fallback chain (smartest → smallest)

fix(frontend): toast icon must be valid emoji, not ✓

fix(frontend): EEG default path points at fixture (HF container has no data/raw/eeg.fif)

feat(frontend): premium motion layer — Apple HIG / Netflix transitions

feat(frontend): editorial redesign — neutral-gray dark + sand accent + dual theme

feat(frontend): Experiments tab — MLflow runs table + two-run diff

feat(api): GET /experiments/runs + POST /experiments/diff (Track 5)

feat(frontend): inline AI Assistant in EEG + MRI tabs

feat(api): POST /explain/eeg + /explain/mri — full-stack Track-1 coverage

feat(llm): modality dispatch — explain(payload, modality) for BBB/EEG/MRI

feat(frontend): AI Assistant tab — natural-language explainer

feat(api): POST /explain/bbb — natural-language rationale endpoint

feat(llm): explainer with deterministic template + OpenRouter fallback

fix(schemas): silence Pydantic protected-namespace warning on ModelProvenance

feat(api+frontend): MLflow provenance badge in decision card

feat(frontend): drift metric line + last-prediction session state

feat(api): drift z-score in /predict/bbb response

feat(models): train-time confidence stats stashed on _neurobridge_train_stats

feat(frontend): MRI tab — Pre/Post ComBat KDE + site-gap KPI strip

feat(api): POST /pipeline/mri/diagnostics — pre/post ComBat KPIs + long-format rows

feat(pipelines): compute_harmonization_diagnostics — long-format pre/post ComBat for viz

feat(frontend): trust caption — precision-at-confidence below decision card