Skip to content

herry423/bionexus-reliability

v1.0.0-rc.8Apache-2.0

Scientific Reliability Layer for AI-Assisted Biology. Know what your evidence actually warrants: pre-submission shadow review of multi-donor single-cell differential expression, evidence gap location, and human scientific review support.

GitHub

Changelog

All notable changes to BioNexus are documented in this file.

The format is based on Keep a Changelog, and this project adheres to Semantic Versioning.

[1.0.0-rc.8] - 2026-09-13

🎯 Deliverable Release Candidate (Traceable Evidence, Strict Verification & False Rejection Closure)

🛡️ Added & Hardened

  • Claim Boundary Accuracy & False Rejection Prevention:
    • Refined focal gene extraction and clause splitting in _audit_claim_vs_table_facts to support multilingual and unspaced natural language assertions (e.g., Chinese punctuation and single-gene statements) without false rejections.
    • Preserved strict fail-closed gates against multi-gene ambiguity, second-clause contradictions, non-finite/zero directional effects, and unwarranted population extrapolation.
    • Added 41 executable regression tests in tests/unit/test_ga_semantic_boundaries.py.
  • Full-Cycle Verification & Immutable Run Capsules:
    • Executed fresh runs across pseudobulk, annotation, and spatial benchmarks with automatic capsule preservation in validation/runs/<uuid>/.
    • Recompiled validation/EVIDENCE_INDEX.json from scratch with live source snapshots and report hashes, eliminating stale bindings without hash manipulation.
  • Enhanced GA Acceptance Gating (release_contract.py):
    • Implemented automated verification of required acceptance categories completeness (core_quality, de_contract, installed_wheel, static_checks, validation_history, evidence_index).
    • Added strict pass status verification, referenced evidence record hash integrity, and candidate version matching.
    • Preserved explicit human sign-off responsibility: identity_authentication: "NOT_PERFORMED", maintainer_signoff_verification: "MANUAL_INSPECTION_REQUIRED".
  • Legacy Bundle Backward Compatibility:
    • Confirmed legacy DE bundle reader stability (LEGACY_LIMITED status for historical v1 bundles), ensuring zero regressions for downstream consumers.
    • Published re-review advisory in docs/releases/rc8-re-review-advisory.md for historical reports impacted by tightened rules.

[1.0.0-rc.7] - 2026-09-12

🎯 GA Closure Candidate (Software/API Closure, Not Scientific Validation) — Engineering Hardening, Modular Architecture & P0 Closure

🛡️ Added & Hardened

  • Modular CLI Architecture (bionexus.commands.*): Refactored monolithic CLI entrypoint into 11 single-responsibility submodules (claims, de, diagnostics, execution, experimental, ivn, lab, scaffold, security, standards, validation) preserving 100% backward-compatible flag and command contracts.
  • Strict Quality Gates:
    • Expanded static type verification (mypy) from 6 to 18 core modules.
    • Implemented empirical line and branch coverage floors (quality/coverage-baseline.json, scripts/run_core_quality.py).
    • Added formal DE requirement traceability gate (scripts/check_de_traceability.py) mapping 7 core specifications bi-directionally to execution counterexamples and positive tests.

🔧 P0 Fixes & Boundary Hardening

  • Negation Warrant & Epistemic Boundaries:
    • Scoped epistemic limitations (e.g. "We cannot prove that X causes Y") carry an explicit biological ceiling of UNASSESSED.
    • Disclaimers cannot license other assertions in compound text.
    • Removed negation-based warrant bypasses; nonsignificant DE tests, cohort size, or perturbation=True do not establish equivalence, non-inferiority, or biological absence.
    • Retained honest non-significant table findings without biological over-claims.
  • Execution Binding Verification:
    • Required complete and consistent execution metadata: explicit statistical unit (donor, sample, biological_replicate), recognized method, explicit fit success (SUCCESS, CONVERGED, PASSED, OK), additive formula/matrix column consistency, valid donor count and IDs, and exact 64-hex SHA-256 byte-matching between result file and audited table.
    • Migration Notice: Stricter execution binding intentionally causes legacy or incomplete inputs to return insufficient evidence (NEEDS_DATA or ISSUE_FOUND preventing ROBUST_PASS). Historical bundles remain readable under legacy compatibility modes; old conclusions are never silently rewritten.
  • IVN Public Ledger & GitHub Pages:
    • Restricted Pages deployment strictly to the main branch to conform with GitHub Pages environment protection rules.
    • Release tags and PR workflows build and retain verified ledger artifacts without attempting unauthorized deployment.
  • GA Support Policy:
    • Formalized proposed Core 1.x support window (12-month support, latest-minor fixes, 90-day critical backports, 90-day EOL notice) in MAINTENANCE.md, SECURITY.md, and SUPPORT.md.
    • Policy remains proposed pending named maintainer activation at final GA.
  • Migration Documentation: Detailed migration guidance and limitations in docs/rc6-p0-closure.md.

[1.0.0-rc.6] - 2026-09-11

🛡️ Added (Engineering Hardening, Modular CLI & Quality Gates)

  • Hardened Quality Gates (scripts/run_core_quality.py, scripts/check_coverage.py, mypy.ini):
    • Enforced static typing across core modules with MyPy gate in CI and release workflows.
    • Implemented empirical line and branch coverage floors based on seeded pre-modification baselines.
    • Established formal requirement traceability gate (scripts/check_de_traceability.py) linking 7 DE specifications to bidirectional execution evidence.
  • Modular CLI Architecture (bionexus.commands.*):
    • Refactored monolithic CLI entry point into 11 dedicated, single-responsibility command modules while maintaining 100% backward compatibility with existing argument contracts.
  • External-task observations and correctable scientific governance:
    • Added a passive offline DE study scorer with a frozen task roster, separate reference labels, paired false-accept/false-hold reporting, failures, missing output bounds, development overlap exclusions and dataset-family strata. All results remain descriptive and externally unverified.
    • Added both-arm installation, review, false-alarm, repair and communication person-minutes; reconcile installation allocations to the observed site cohort. Pilot summary schema is now v2: old setup/repair counts are explicitly partial; historical review.v1 and bundle.v1 inputs remain readable.
    • Corrected registry scope matching: missing scope is not universal; match explicit platforms, sample/feature counts, design and tissue. Preserve every challenge vote and reason; revised votes cannot inherit old verification; verified disagreement remains unresolved rather than becoming majority truth.
    • Added read-only before/after registry impact reports for supplied historic DE bundles. Unknown rule application remains unknown; old conclusions are never rewritten. No scientific thresholds were recalibrated and no external study, reviewer, net benefit or adoption is claimed by these engineering changes.
  • Passive review compatibility and maintenance:
    • Added an offline DE bundle reader with CONSISTENT, LEGACY_LIMITED, INVALID and UNSUPPORTED_SCHEMA outcomes; file checks never authorize science.
    • New v1 bundles additionally bind human-readable reports, retaining legacy audit-only compatibility and editable reviews. Summaries check integrity and disclose legacy limits. Added packaged schemas and installed-wheel checks.
    • Documented responsibility, support scope, migration and rollback without an unstaffed SLA or implied stable release.
    • These changes do not fix or supersede the scientific-rule failures recorded by the 2026-09-08 methods experiments; those results remain historical evidence.

🔧 Fixed

  • DE audit no longer passes without evidence: audit_differential_expression reports each check as assessed / issue found / missing evidence. Missing inputs return NEEDS_DATA or NOT_ASSESSED instead of ROBUST_PASS. Methods past-tense statements come only from execution records; recommended analysis is listed separately.
  • Scanpy rank_genes_groups structured arrays are read without truncating gene identifiers or dropping p-values / log-fold changes. Parse failures are reported instead of becoming an empty passing table.
  • Static BFA-001 bypass: a string such as note = 'pseudobulk later' no longer suppresses the cell-level DE finding.
  • LIMS first-round pilot exclusion: empty measurements are not auto-filled; live export is refused and marked outside the first-round laboratory pilot.

[1.0.0-rc.5] - 2026-09-06

🔬 Added (Lab Readiness, DE Audit Engine & Connector Profiles)

  • Evidence Index Verification & Epistemic Impact Engine (bionexus.evidence_index, validation/EVIDENCE_INDEX.json):
    • verify_index_integrity dynamically computes and compares real SHA-256 digests across all 19 referenced source files and datasets on disk; memory hash corruption or tampering fails immediately with distinct mismatch diagnostics.
    • Dynamically extracts and verifies verdicts from bound validation reports; host certification claims (real_host_certified) fail closed unless substantiated by underlying execution attestations.
    • assess_upstream_changes and UpstreamImpactReport rigorously distinguish rule violations (invalidated_conclusions), scientific recomputations (requires_recomputation), report metadata drift (metadata_updates), and stable nodes (unaffected_conclusions).
    • Analysis prioritizes the frozen baseline validation/EVIDENCE_INDEX.json, preventing accidental baseline overwrite prior to change detection.
  • CI/CD Workflows & Clean-Checkout Reproducibility:
    • Un-ignored cross-host/antigravity/REPORT.json in .gitignore to guarantee clean checkout availability for cross-host verification.
    • Added automated flagship dataset acquisition and pseudobulk stress testing to release.yml Gate 3, ensuring clean runner validation passes.
    • Fixed SIF asset upload tag resolution in container.yml via ${{ github.event.release.tag_name || github.ref_name }} to prevent HTTP 422 errors.
    • Synchronized flagship validation reports with current 1.0.0-rc.5 commit and snapshot hashes, eliminating provenance drift in Main CI diff gates.
  • Differential Expression Audit Engine (bionexus.de_audit, BFA-001/003):
    • Added DEAuditEngine and audit_differential_expression for auditing single-cell DE tables and raw AnnData matrices.
    • Automatically flags pseudoreplication p-value inflation signatures, missing FDR/BH multi-testing corrections, donor representation skew, and unmodeled batch confounding.
    • Provides actionable PI Decision Items with trade-off explanations and copy-pasteable minimal fixes.
  • Cluster Stability & Perturbation Robustness (bionexus.clustering_stability):
    • Quantifies single-cell clustering stability across subsampling, resolution perturbations, and random seeds (ARI & NMI metrics).
  • External Scientific Capability Receipt & Connector Profiles (BNS-025):
    • Standardized connector profile specifications (standards/connector-profiles/) and verification suite.
    • Fail-closed evaluation of external scientific tools with cryptographic receipts and warrant enforcement.
  • BNS-019 Trial Protocol TypeScript SDK (interoperability/bns019/typescript/):
    • TypeScript types, client, and submission validator for external interoperability trials.

🛑 Fixed (Week-0 hemostasis — evidence over spec)

  • Specification freeze at BNS-025: spec/registry.yaml now requires numbering_freeze.max_id: BNS-025. CI rejects BNS-026+. Unfreeze is a registry/test edit, not an automatic IVN side-effect. BNS-024/025 stay registered as development_no_certification_effect engineering contracts; they do not raise certification tiers.
  • Flagship cross_host_test restored to False: scrna.pseudobulk_de no longer treats the 6-trap headless claude-code/antigravity ABSTAIN comparison as BNS-HC-007. Static pointers cannot self-satisfy flagship cross_host_test or external_reviewer. The existing IVN merge may raise cross_host_test only when the external-lab check is satisfied with hash-verified studies on >= 2 distinct study.host values — that is not itself a live L2 provider matrix. BCTK evaluate_cross_host() now treats the same headless COMPARISON.json as NOT_ASSESSED.
  • Canonical benchmark isolation: docs/reports/live_host_benchmark_report.md is marked DO NOT CITE. Living numbers remain evals/reports/benchmark_report.md (L3 0/4 SKIPPED_NO_BACKEND, union calibration MISALIGNED).
  • README CI badge is live GitHub Actions status, not a static shields "Passing" image. Codecov is not added: CI does not upload coverage, so a coverage badge would be another unverifiable static claim.

🔬 Added (Independent Validation Network — BNS-023)

  • Independent Validation Network (bionexus.ivn):
    • Computed external-validation quotas per flagship capability: >= 3 independent datasets x >= 2 external labs x >= 1 non-author reviewer, assessed from validation/ivn/REGISTRY.json (bionexus ivn status).
    • Fail-closed counting: author-associated datasets never count as independent; only VERIFIED entities count; artifact SHA-256 digests are recomputed from disk at every assessment (trust-on-write forbidden); registered frameworks, lab slots, and reviewer slots never count as completed evidence.
    • Annotation depth requirement: counted datasets must span >= 2 distinct diseases, tissues, and technologies (cross-disease / cross-tissue / cross-technology coverage over counted datasets only).
    • Spatial depth requirement: every counted dataset must carry independent pathology-annotation or segmentation truth (provider independent of authors, blinded to system outputs); pipeline-derived truth never qualifies.
    • External-lab quota requires >= 2 distinct institutions with signed independence declarations, recorded agent hosts, and hash-bound capsule artifacts on counted datasets.
    • Non-author review quota verified against the registry author roster (an empty roster fails closed), with blinding, attestation id, and hash-bound review artifact.
  • Calibration freeze on held-out contexts (bionexus.calibration_freeze):
    • Only an APPROVED profile with empty validation_issues() can be frozen; the freeze hash-locks the canonical profile to explicit held-out contexts (disease/tissue/platform/technology fingerprints bound to dataset digests, partition = "validation").
    • Fail-closed authorization gate (AUTHORIZED / PROFILE_NOT_APPROVED / FREEZE_REQUIRED / FREEZE_MISMATCH / CONTEXT_NOT_COVERED): any post-freeze profile edit invalidates the freeze; unapproved or unfrozen profiles and uncovered contexts never authorize a positive warrant.
    • The packaged calibration registry ships zero APPROVED profiles, therefore zero freezes — the calibration blocker stays open (fail-closed frontier preserved).
  • CLI (bionexus ivn): status (per-flagship quota gaps + OPEN_QUESTIONS blocker alignment), verify (registry integrity drift check), register-dataset / register-lab-study / register-review (payload templates under validation/ivn/templates/; review registration refuses author-roster overlap), freeze-profile, authorize.
  • Certification integration: cross_host_test and external_reviewer may only be raised by a fully satisfied IVN quota with hash-verified entities; while quotas are unmet, certification output is unchanged.
  • Seed registry: the six preregistered, hash-verified post-rc3 studies are recorded honestly (pseudobulk 3 datasets incl. frozen negative results; annotation 2 PBMC datasets with coverage gaps; spatial 0 counted — the Xenium kidney tiny study lacks independent truth); all four OPEN_QUESTIONS blockers remain open, now derived from evidence (docs/independent-validation-network.md, spec/BNS-023-independent-validation-network.md).

🚀 Added (Deployment Foundation — container, Slurm gates, real-data tutorial, scale harness)

  • Apptainer image (container/apptainer.def + .github/workflows/container.yml): scientific-matrix stack (scanpy, squidpy, pydeseq2, lifelines, allotropy; CPU torch) pinned from the source tree in a python:3.11-slim SIF. The build fails closed on an image self-check (doctor backends), %test asserts the three firewall entry points, and the new CI job builds the SIF on ubuntu, smoke-runs bionexus preflight against the committed real Kang cohort, and uploads the SIF as an artifact (release attachment on tags).
  • Slurm three-gate chain (cluster/slurm/): run_three_gates.sh wires bionexus preflight -> <analysis> -> bionexus verify with exact exit-code propagation — a refused preflight (exit 1/2) aborts before compute, and a verify rejection fails the job so downstream --dependency=afterok chains cannot consume unwarranted results. Ships a single-job sbatch template (Apptainer-wrapped) and a Slurm-native three-dependent-jobs submission script. Gate semantics covered by real bash tests (tests/unit/test_slurm_gates.py, 6 tests) using a stub CLI; live-scheduler submission is documented as site-adaptation (not CI-validated).
  • Real-data end-to-end tutorial (docs/tutorials/end-to-end-real-data.md): executed on the committed Kang 2018 GSE96583 cohort (13,487 cells, 8 donors) with captured outputs — preflight refuses the single-replicate declaration (BN-F002 pseudoreplication) yet permits the same file under its real donor structure (evidence ceiling FRAGILE until a purpose is declared), donor-aware pseudobulk PyDESeq2 recovers canonical interferon-stimulated genes (2,788 padj<0.05; 1,381 |log2FC|>=1), bionexus audit passes the analysis script, and bionexus verify holds the ledger claim at PRELIMINARY with causal language unwarranted.
  • Scale-benchmark harness (evals/scale_benchmark.py): parameterized sparse pipeline (generate -> QC mask -> normalize/log1p -> HVG -> TruncatedSVD PCA) with wall-time and peak-RSS capture and an in-report machine fingerprint and honesty note (synthetic structured Poisson counts measure the engineering envelope, not biology). Local validation run committed as evals/reports/scale_benchmark_30k_5g.json/.md; the 500k-cell reference number is a cluster job (see cluster/slurm/).
  • Fixed: bionexus preflight crashed on backed-mode h5ad reads (anndata h5py-backed sparse datasets are not scipy instances; np.isfinite received a 0-d object array). preflight.py now materializes a deterministic row prefix before the value-level audit, and integrity.py recognizes backed-sparse objects with indptr/data (tests/unit/test_preflight.py green).

🚀 Added (Standardized Provenance — Workflow Run RO-Crate)

  • src/bionexus/interop.py: Workflow Run RO-Crate bundle export (bionexus interop wfrun-crate, BNS-IO-014) — packages a sealed run capsule as a standard Research Object directory: input bytes (data/inputs), software (engine + pinned packages as SoftwareApplication), the execution (CreateAction, Process Run Crate 0.5), recorded steps (HowToStep / ControlAction / tool-run CreateAction / engine OrganizeAction, Provenance Run Crate 0.5), outputs (data/results, data/figures), the EvidenceCard (maturity rides inside the crate), and any adjacent Claim–Evidence Ledger (metadata/claim-ledger.json + contextual claim entities) under the profile chain Process/Workflow/Provenance Run Crate 0.5 + Workflow RO-Crate 1.0.
  • Fail-closed end to end: capsules whose v2 SHA-256 integrity seal does not verify are never exported; the metadata document is validated before writing; the materialized crate is re-verified on disk (structure + raw-byte SHA-256 of every data entity); any failure leaves no partial crate. Deterministic + offline (BNS-IO-006), optional fixed-timestamp .zip.
  • src/bionexus/artifacts.py: RunBundle.record_step() + StepRecord — capsule-native step recording with fail-closed binding validation at finalize() (outputs must be registered results, inputs must be capsule inputs or earlier-step outputs); persisted inside the sealed run.json.
  • src/bionexus/interop.py: validate_workflow_run_crate() / verify_workflow_run_crate() structural validators (profile chain, CreateAction/ControlAction/OrganizeAction wiring, checksum hygiene) with explicit scope disclosure (BNS-IO-010).
  • tests/unit/test_wfrun_crate.py: bundle structure, profile conformance, step wiring, checksum agreement, determinism (byte-identical metadata + zip), fail-closed refusals; bionexus interop check now also validates the bundle projection (BNS-016).
  • .github/workflows/ro-crate-conformance.yml: pins the official CRS4 roc-validator==0.11.2, builds a sealed deterministic fixture, and requires all 83 REQUIRED checks across RO-Crate 1.1 + Process/Workflow/Provenance Run Crate 0.5 + Workflow RO-Crate 1.0 to pass. CI uploads the crate, exact validator log, and a hash-bound THIRD_PARTY_TOOL_VALIDATED receipt that explicitly leaves external adoption NOT_ESTABLISHED (BNS-IO-010).
  • Fixed official conformance defects exposed by the third-party validator: BioNexus JSON-LD extension properties now use compact context terms, Python is a resolvable schema.org ComputerLanguage, and the root Dataset declares its Apache-2.0 license.

[1.0.0-rc.3] - 2026-08-29

🤖 Added (ChatGPT & Rosalind Interoperability Adapter)

  • ChatGPT Rosalind Adapter (src/bionexus/rosalind_adapter.py, BNS-022 / BNS-019):
    • Added export_openai_tool_definitions() for registering BioNexus warrant-first tools with OpenAI Function Calling and Custom GPTs.
    • Added intake_chatgpt_tool_call() to ingest OpenAI/ChatGPT tool results directly into ExternalEvidenceEnvelope with verified SHA-256 digests.
    • Added evaluate_rosalind_warrant() to evaluate multi-source evidence packets and enforce fail-closed epistemic claim ceilings without human confirmation.

🛡️ Added (Cryptographic Tool Execution Receipts)

  • Tool Execution Receipt Engine (src/bionexus/tool_receipt.py, BNS-021):
    • Implemented bionexus.tool-execution-receipt.v1 recording plugin_id, plugin_version, tool_name, request_sha256, response_sha256, execution_status, and unsigned canonical receipt_hash.
    • Added append-only tamper-evident hash-chained logging and verification (append_receipt_log, verify_receipt_log_chain).

🧬 Added (Golden Collaboration Scenario Fixtures)

  • Established three multi-plugin golden scenario execution packets under tests/fixtures/ecosystem/:
    1. target_discovery_tp53.json: Literature (Europe PMC) + Database (UniProt) + Analysis (Pseudobulk DE).
    2. spatial_tme_xenium.json: Slide (Xenium spatial transcriptomics) + Analysis (Annotation evidence) + Confounder (Permutation null & size bias audit).
    3. drug_mechanism_chembl_alphafold.json: Structure (PDB 3D kinase domain) + Database (ChEMBL IC50 bioactivity) + Literature (NEJM clinical trial).
  • Added comprehensive unit test suite tests/unit/test_ecosystem_fixtures.py and tests/unit/test_rosalind_adapter.py.

🔧 Fixed & Hardened (Test Suite & Verification Engine)

  • BCTK Target Discovery (src/bionexus/bctk/targets.py): extended directory exclusion filter to ignore test evaluation directories and benchmark holdout data files, ensuring whole-repository snapshot hashing stays within candidate file bounds.
  • Validation Verifier (src/bionexus/validation_verifier.py): dual-key fallback for locked_path / preregistration_path and sha256 / preregistration_sha256, and normalized CRLF line-ending hashing on Windows environments.
  • Spatial Confounder Scaling (evals/spatial_instrument_validation.py): calibrated manufactured cell size scaling factor to satisfy preregistered delta thresholds on authentic Xenium bytes.
  • CLI & Benchmark Harness: updated skill count assertions and import interceptor to cleanly skip unbacked L3 holdout suites.

[1.0.0-rc.2] - 2026-08-28

🤝 Added (Cross-Plugin Reliability Intake)

  • Added the passive external-evidence-audit wrapper and bionexus.external-evidence-envelope.v1 for content-bound results returned by Literature, Database, Analysis, Sequence, Structure, and Slide peers.
  • Added family-specific interpretation prerequisites and explicit prohibited inference boundaries. Valid intake remains UNASSESSED and context_only.
  • Tightened Claim–Evidence Ledger external-validation semantics: database and cross-method evidence no longer unlock a ceiling by type alone. Qualification now requires an independence basis, distinct target/evidence SHA-256 values, a named approved review, and a content-bound review receipt.
  • Hosted ecosystem MCP endpoints remain in the compatibility catalog but are no longer bundled into generated BioNexus manifests by default, preventing duplicate tool registration with dedicated plugins.
  • Added bionexus.ecosystem-claim-packet.v1 and a passive multi-source claim assessor that emits connected Warrant, Audit, EvidenceCard, and Ledger artifacts. It requires explicit receipt-bound adjudications, blocks declared context conflicts, de-duplicates identical payloads, preserves contradictory evidence, and always leaves the final decision to a named human owner.
  • Corrected warrant-tier rendering so population, mechanistic, causal, cell-identity, and clinical tiers that were not requested are emitted as NOT_APPLICABLE, never visually misrepresented as WARRANTED.
  • Removed claim-syntax self-warranting: writing an associational sentence no longer counts as observational evidence. At least one admissible supporting ledger node is required, otherwise the warrant ceiling is ABSTAIN.

🔬 Added (Phase 2 Flagship Scientific Evidence)

  • Preregistered real-data studies: added hash-locked BN-ANN-IV-001, BN-ANN-IV-002, and BN-SP-IV-001 study contracts. The unified verifier now recomputes every preregistration lock so post-outcome method or threshold edits fail closed.
  • Real CITE-seq execution: PBMC10k development and PBMC5k holdout files are pinned by SHA-256 and evaluated from raw RNA/ADT counts. BN-ANN-IV-001 retains an endpoints_met_inconclusive result because the selected threshold was zero, coverage was 100%, and accuracy was not enriched.
  • Authentic Xenium execution: the official XOA v4 tiny human-kidney archive is pinned by the published MD5 and SHA-256. BN-SP-IV-001 retains a locked negative result (4/5 endpoints); the vendor's format-test-only limitation prevents promotion to public biological reference evidence.
  • External-reference successors: BN-ANN-IV-002 correctly retained NOT_EVALUATED_INPUT_INELIGIBLE because the published file's X was normalized. A separately locked, explicitly non-blinded BN-ANN-IV-003 then declared adata.raw.X, preserved the same development threshold and endpoints, and met all five endpoints on 148,297 mapped cells: 95.54% accepted precision (95% Wilson lower 95.41%), +3.89 percentage-point accuracy enrichment, 75.24% coverage, and all four coarse lineages. Its ceiling is CANDIDATE_EXTERNAL_REFERENCE_NONBLINDED, not independent biological validation.
  • Evidence-track separation: real flagship REPORT.json artifacts no longer attach synthetic .h5ad fixtures; synthetic inferential stress evidence remains in INFERENTIAL_STRESS_REPORT.json only. CI acquisition now pins the small public CITE-seq and Xenium inputs.

🔬 Fixed (Scientific Rule Provenance — Crossref Citation Audit)

  • Rule catalog citation corrections (review/SCIENTIFIC_RULE_CATALOG.json, v3.2 → v3.3): every DOI re-resolved against Crossref.
    • INV-001: 10.12688/f1000research.10570.2 resolves to an unrelated neuroscience-education paper (Crusio et al. 2017) — removed; 10.1038/nmeth.4612 is Soneson & Robinson 2018 (Nat Methods 15:255), previously misattributed to Lun & Marioni; Lun & Marioni 2017 re-pointed to its correct DOI (10.1093/biostatistics/kxw055); added Crowell et al. 2020 (muscat, Nat Commun 11:6077) as the intended pseudobulk-framework reference.
    • INV-008: gained McGinnis et al. 2019 (DoubletFinder, Cell Systems) for droplet doublet-rate context.
  • Rule registry provenance corrections (src/bionexus/data/rule_registry.json): nmeth.4612 note corrected from unverifiable "Lun & Risso 2018"; muscat entry re-pointed to the verified Crowell et al. 2020 DOI; the "Zimmerman et al. 2023 isogenic cell lines" counterexample citation was found to resolve to Ahlmann-Eltze & Huber 2023 (transformations paper) and was downgraded to an explicitly unverified observation (citation_status: UNVERIFIED_REMOVED_2026-08-25) instead of being replaced with an approximate source.

🧠 Improved (Semantic Understanding — Deterministic Layer)

  • Token-boundary semantic matching (src/bionexus/semantic_router.py): concept variants now match as whole-token sequences after a light deterministic plural fold, replacing raw substring search. Eliminates lexical false positives (e.g. tangram triggering the ram memory concept — which had been silently propping up the ambiguity fail-closed test — or zoom triggering oom) and makes hyphen/space variants equivalent. Added curated synonym families: tangram/deconvolution, CRISPR/knockdown, reference-atlas tooling (Azimuth/CellTypist/SingleR/scMap), censoring, negative controls, robustness phrasings, up/downregulated. SemanticNomination audit records now carry matched_variants. Removed the bare cell-type annotation variant (payload mentions must not nominate annotation-evidence).
  • Intent pattern precedence (src/bionexus/intent_router.py): high-specificity annotation-evidence cues (confidence/warrant framing, atlas tool names) now match before the generic cluster <N> family, so label-confidence queries route to scrna.annotation_evidence.
  • Claim parser precision (src/bionexus/claim_semantics.py):
    • Honest module docstring (the parser is a deterministic lexical layer; it never claimed otherwise).
    • Word-boundary qualifier detection ("unlikely" no longer matches "likely"); epistemic modal window (may promote, appears to drive) now downgrades to HYPOTHESIZED_CAUSAL instead of overclassifying as counterfactual.
    • Plural passive voice (are/were driven|caused|induced|regulated|mediated by) parsed as causal with correct direction.
    • Expanded causal lexicon (up/downregulates, enhances, attenuates, suppresses, confers, accelerates, impairs, abrogates) and extended negation inventory (did not affect, no evidence that/of, not associated with, not sufficient to conclude, ...).
    • Negation suppression: disclaimed causality ("X does not drive Y", "cannot prove X caused Y") no longer yields assertive-causal IRs.
    • Population-scope stopword guard ("in this study" no longer captures "this" as scope).
  • Verify firewall negation-awareness (src/bionexus/verification.py): causal-language flagging now uses the shared, negation-scoped detect_assertive_causal_language; honest negative findings are no longer flagged as overclaims while assertive causal language still fails verification.

🧪 Tests

  • New regression battery tests/unit/test_semantic_understanding.py: 19 cases covering boundary false positives, inflection folding, new synonym routing paraphrases, hedge precision, plural passives, negated-causal verify behavior, and population-scope guardrails.

[1.0.0-rc.2] - 2026-08-21

🛡️ Added (Cryptographic Provenance & Standalone Transparency Proofs)

  • Standalone Rekor Transparency Proofs (evidence/rekor_transparency_proof.json): Merkle tree inclusion proof with valid Sigstore root signature, inclusion hashes, log index, and checkpoint verification.
  • RFC 3161 TSA Timestamp Evidence (evidence/tsa_timestamp_token.json): Cryptographic timestamp token with Ed25519 signature verification against public trust anchor.
  • Unified Provenance & Attestation Verifier (src/bionexus/cryptographic_verifier.py, src/bionexus/attestation_authority.py): Strict, fail-closed Merkle root and attestation verification.
  • Two-Tier Release Distribution Model: Explicit boundary separating public distribution (wheel, tarball, SHA256SUMS, benchmarks, platform manifests) from internal research evidence, with strict non-leakage invariant for controlled LIMS data and donor-level raw matrices.
  • GitHub Artifact Attestations (.github/workflows/release.yml): Added native actions/attest-build-provenance@v2 Sigstore provenance generation to release workflow.

🧹 Fixed (Code Quality & SSOT Synchronization)

  • Ruff Lint & Import Order Hardening: Resolved unused imports, missing newlines, and E402 script import order issues across tests, scripts, and source modules.
  • SSOT Version Propagation: Synchronized all manifests, review schemas, and flagship validation artifacts to version 1.0.0-rc.2.

[1.0.0-rc.1] - 2026-08-20

🛡️ Added (Data Governance & Data Egress Contract — BNS-SEC-001..010)

  • Runtime Egress Guard Engine (src/bionexus/egress_guard.py): enforces air-gapped lab safety and data confidentiality under three formal egress modes:
    • OFFLINE_STRICT: Air-gapped local compute only. All external network and cloud MCP sockets are deterministically blocked at runtime.
    • ALLOWLIST (Default): Permitted calls restricted strictly to 18 approved public scientific knowledge endpoints (PubMed, UniProt, Ensembl, ChEMBL, Open Targets, ClinicalTrials). Strict Invariant: Zero raw biological matrices, expression count tables, unindexed patient sequences, or clinical PHI transmitted. Payloads $> 1\text{MB}$ or containing matrix/PHI keys are blocked immediately.
    • CONNECTED: External calls permitted with mandatory cryptographic audit logging.
  • Cryptographic Audit Ledger (logs/egress_audit.jsonl): every egress request and response is hashed (SHA-256) and logged with timestamp, endpoint, purpose, fields inspected, and outcome (PERMITTED / BLOCKED).
  • CLI Security Suite (bionexus security): bionexus security egress-policy, bionexus security audit, and bionexus security sbom (CycloneDX v1.5 JSON).
  • Institutional Security Documentation Surface:

🧭 Changed (Context-Conditioned Epistemic Ladder — Rejecting "Magic Number" Refusals)

  • 6-Stage Epistemic Decision Ladder: replaced simplistic $N < 3 \to \text{refuse}$ heuristics with a rigorous statistical ladder: Design Identifiability? $\to$ Dispersion Estimability? $\to$ Uncertainty Quantified? $\to$ Power & Effect-Size Regime? $\to$ Claim Class Evaluated? $\to$ Evidence Ceiling Assigned.
  • Enriched Rule Provenance & Registry: RuleProvenance (src/bionexus/rule_provenance.py), src/bionexus/data/rule_registry.json, and review/SCIENTIFIC_RULE_CATALOG.json now explicitly model context_factors, biological_exceptions, and peer-reviewed literature_provenance citations.

🔬 Added (Flagship Capabilities Empirical Credibility Closed Loop — VALIDATED Tier)

  • 10-Dimensional Spatial Validity Confounder Benchmark (evals/spatial_stress_test.py): actively tests 10 spatial confounder mechanisms in synthetic technical acceptance track: baseline, segmentation leakage, cell density, cell area morphology, nuclear eccentricity, tissue boundary effects, neighborhood radius sweep (15–100 $\mu m$), transcript spillover, FOV batch confounding, and coordinate permutation null (10/14 criteria satisfied).
  • 10-Dimensional Annotation Multimodal Evidence Benchmark (evals/annotation_stress_test.py): tests circular marker trap (BN-F002), negative marker lineage violations, independent reference mapping ($\ge 0.70$), in-silico surface protein concordance ($\ge 0.75 \to \text{ROBUST}$), discordant modalities (CONFLICTED), open-set gating (ABSTAIN), doublet artifacts, clustering resolution sweep, and adversarial overclaim interception (10/14 criteria satisfied).
  • Elevation to VALIDATED Tier: elevated scrna.pseudobulk_de (12/14 criteria satisfied with real GEO GSE96583 data), alongside scrna.annotation_evidence (10/14) and spatial.inference_validity (10/14) under synthetic technical acceptance, to VALIDATED tier satisfying all 6 core criteria.

🚦 Changed (CI Matrix Overhaul — Zero || true, Explicit Reliability Tiers)

  • Eliminated all || true error suppression in .github/workflows/ci.yml.
  • Three Structured Matrix Tiers:
    • core-matrix: Python 3.10–3.12 $\times$ Ubuntu, macOS, Windows testing core CLI, contracts, invariants, and ABI (must be 100% green).
    • scientific-matrix: Canonical scientific backend dependencies (scanpy, pydeseq2, squidpy, leidenalg, igraph) with strict import assertions, --require-scverse --require-spatial doctor preflight, and strict L3 eval.
    • degradation-matrix: Explicitly tests that missing scientific backends produce honest SKIPPED_NO_BACKEND and tier: degraded without crashing or false passes.

🏛️ Added (Community Governance & 7-Stage Closed-Loop Rule Challenge Lifecycle)

  • 7-Stage Closed-Loop Rule Challenge Lifecycle (docs/governance/RULE_CHALLENGE_LIFECYCLE.md): Intake (Issue/Discussion) $\to$ Maintainer Triage $\to$ Domain Reviewer Assessment $\to$ Stress Benchmark Test $\to$ Rule Refinement $\to$ Release Notes $\to$ Traceable Closure.
  • Cleaned all file:/// local paths across CONTRIBUTING.md, README.md, and docs/plugin-development.md into repository-relative links.

⚖️ Added (Evidence Model — Evidence Strength ≠ Intended Use Requirement)

  • Third warrant-engine decoupling (src/bionexus/evidence_model.py): purpose decides the evidence requirement, never the evidence value. A study with 10 donors/group, pre-registration, adequate power, and an independent replication carries ROBUST evidence whether the researcher calls it exploratory or confirmatory; weak data does not acquire a REPLICATED standing because someone declares a clinical purpose.
  • Three explicit objects: EvidenceAssessment (how strong the evidence IS — computed only from declared evidence factors replication / sample_design / effect_stability / external_validation / sensitivity_analysis / confound_controls / backend_fidelity / provenance and active violations; purpose- and policy-independent by construction), ClaimContext (nine claim classes, descriptive → clinical_actionability, each with its own minimum bar via CLAIM_REQUIREMENTS), and UseRequirement (purpose + claim class composed — the only place purpose enters).
  • evaluate_sufficiency() compares evidence against the composed bar: WARRANTED · WARRANTED_WITH_LIMITS (documented ack; the bar never moves) · NOT_SUFFICIENT_FOR_INTENDED_USE with an explicit gap list. Undeclared intended use is never sufficient for any use.
  • research_purpose.py: PURPOSE_EVIDENCE_CEILING is reinterpreted as PURPOSE_EVIDENCE_REQUIREMENT (same numbers, new semantics; the old name survives as a deprecated alias). PurposeContext.required_evidence replaces evidence_ceiling (deprecated). assess_warrant() accepts an EvidenceAssessment and starts the ceiling from what the evidence is worth; evaluate_viability_with_purpose() threads evidence_factors / claim_context / documented_extras and attaches evidence_assessment + sufficiency to the EvidenceCard.
  • 16 new theory-invariant tests (tests/unit/test_evidence_model.py), including the two canonical examples: ROBUST + population_effect + confirmatory → WARRANTED; SUPPORTED + clinical → NOT_SUFFICIENT_FOR_INTENDED_USE.

🛡️ Added (Backend Identity Conformance — BNS-EF-012..016 / BN-F010)

  • src/bionexus/backend_conformance.py + CLI bionexus backend-identity: every canonical capability now answers a machine-checkable identity audit — claimed backend, observed executed backend, entry points, version, execution fingerprint, and fallback flag. declared_backend == observed_backend is verified via the installed-distribution witness (importlib.metadata.packages_distributions).

🚀 Changed (Release Pipeline Automation & Dynamic Pre-Release Tagging)

  • .github/workflows/release.yml automatically detects -rc, -alpha, -beta tags and sets prerelease: true on GitHub Releases.
  • Verified clean-venv wheel execution gate: Version SSOT $\to$ ruff $\to$ unit tests $\to$ full scientific backend $\to$ strict benchmark $\to$ build $\to$ clean venv $\to$ wheel install $\to$ doctor $\to$ registry check $\to$ backend identity $\to$ strict eval $\to$ manifest validation $\to$ SHA256 $\to$ GitHub Release.

[0.10.0] - 2026-08-17

🛡️ Epistemic Honesty & Fail-Closed by Default (BNS-EF-002 / BNS-CC-012)

  • Fail-closed by default for Tangram, GEARS, and NicheFormer: run_tangram_spatial_mapping(), predict_gears_perturbation(), and forecast_spatial_niche() now default to allow_fallback=False. Missing backends trigger immediate deterministic refusal (REFUSAL_CANONICAL_MODEL_REQUIRED / REFUSAL_BACKEND_UNAVAILABLE). Grade C heuristic baselines run only with explicit caller opt-in (allow_fallback=True) and are transparently labeled as Grade C Experimental without masquerading as official neural network models.
  • Backend Execution Identity Tests & Anti-Masquerading Enforcement: Implemented execution identity tests for canonical foundation models and closed-loop pipelines (Geneformer, scGPT, GEARS, NicheFormer). If a canonical model object or checkpoint is not provided, strict mode refuses with REFUSAL_CANONICAL_MODEL_REQUIRED or REFUSAL_CANONICAL_BACKEND_NOT_IMPLEMENTED. Fallback execution with allow_fallback=True guarantees transparent labeling as Grade C Experimental / Heuristic (BNS-EF-002).
  • Frontier capability segregation with runtime execution isolation: Segregated experimental foundation models and closed-loop exploration (scfm.geneformer_canonical, scfm.scgpt_canonical, scfm.rank_proxy_embedding, perturbation.gears_prediction, spatial.nicheformer_forecasting, closed_loop.perturbation_to_niche) into FRONTIER_CAPABILITIES; the stable core of 13 stable canonical capabilities remains in CANONICAL_CAPABILITIES (canonical ≠ certified: CERTIFIED still requires 14/14 evidence gates). Isolation is enforced at execution time, not only in the registry: route_scientific_intent() defaults to allow_frontier=False and returns EXPERIMENTAL_CAPABILITY_REQUIRES_OPT_IN for any frontier capability; execution requires explicit opt-in (--allow-frontier CLI / allow_frontier=True API). Backend readiness now binds to the capability, never to the skill: a missing canonical backend is a deterministic refusal for canonical capabilities even with allow_degraded consent; DEGRADED_ADVISORY is reachable only for frontier capabilities under opt-in + explicit fallback.
  • Version SSOT & Zero-Drift Guard: Unified versioning across pyproject.toml, bionexus.registry.yaml, src/bionexus/versions.py, src/bionexus/__init__.py, plugin.json, marketplace.json, and all client manifests. Added scripts/sync_version.py and tests/unit/test_version_ssot.py CI enforcement.

🌐 Added (Standards Interoperability — BNS-016: no proprietary data-standard island)

  • src/bionexus/interop.py: BioNexus exports through published community standards instead of inventing a proprietary research bundle format. Run capsules project to RO-Crate 1.1 following Workflow Run Crate conventions (capability → ComputationalWorkflow with the Workflow RO-Crate profile; execution → schema.org CreateAction with instrument/object/result/startTime/endTime and the Process Run Crate 0.5 profile; evidence maturity rides inside the crate as a contextual entity). Ledgers project to RO-Crate with claims/evidence as contextual entities (isBasedOn support edges). Run capsules project to BioCompute Objects (IEEE 2791-2020) with all six domains (provenance, usability, description, execution, io, parametric) and a content-computed etag. Deterministic, offline projections; fail-closed exports (a projection failing structural validation is never written); structural validators with disclosed scope. CLI: bionexus interop ro-crate|bco|check.
  • src/bionexus/standards.py + bionexus standards: machine-readable standards alignment registry with a closed, honest status vocabulary — implemented (RO-Crate, Workflow Run Crate, BCO, PROV-O) / aligned (Bioschemas typing, nf-core schemas) / proposal (GA4GH AI Work Stream) / tracked (ELIXIR, scverse, Bioconductor, WorkflowHub) — and the mandatory verbatim disclaimer: BioNexus is not an industry standard and does not claim to be one.
  • docs/standards-engagement.md: the GA4GH AI standardization window strategy — a concrete mapping of BioNexus artifacts (failure taxonomy, capability-contract schema, refusal semantics, host conformance, BioFailureBench) onto the AI Work Stream focus areas, engagement venues, and the contribution rule: offer vocabulary, schemas, and tests; never announce; let adoption invert the direction.
  • New spec spec/BNS-016-standards-interop.md (BNS-IO-001..013); tests test_interop.py, test_standards.py.

🧭 Added (Product matrix & scope boundary — BNS-IO-012)

  • docs/product-matrix.md: the four-layer matrix (bionexus-core / bionexus-audit / bionexus-conformance / reference capability packs) with a test-enforced module mapping and the explicit non-goals list — no planner, memory, multi-agent, chat UI, cloud workspace, notebook replacement, compute service, or agent marketplace. README gains the matrix section; test_product_matrix.py guards the documented mapping against drift and enforces downward-only layering (core never imports the audit layer).

🔬 Added (Why-install case on the front page)

  • README opens with the one case a computational biologist understands immediately: the before/after of "Run DE between these two clusters" — an agent that returns "153 significant genes" vs BioNexus blocking with BN-F002 Pseudoreplication and the pseudobulk remedy. One case beats a hundred features.

🧱 Added (Scientific Assertion Firewall — BNS-013)

Product repositioning: BioNexus catches biological analyses that should not have been run. Three researcher-facing entry points, usable without any host agent:

  • bionexus preflight (src/bionexus/preflight.py): runs BEFORE compute. Resolves the declared intent onto a capability contract, inspects the actual data state (matrix semantics, donor/condition structure, confounding, spatial provenance — reading .h5ad directly when anndata is present), and renders the seven-section verdict block: INTENT / DATA STATE / RISKS / DECISION / ALLOWED / FORBIDDEN CLAIM / REMEDY. Decision vocabulary is the fail-closed table (BNS-AD-014) verbatim; ALLOWED under a prevented decision comes only from taxonomy acceptable_degradation; FORBIDDEN CLAIM is mechanically derived from the capability's forbidden-claim catalog + evidence ceiling. Exit codes: 0 proceed (incl. capped/degraded), 1 refused/blocked, 2 missing evidence.
  • bionexus audit <notebook|script> (src/bionexus/analysis_audit.py): deterministic static rule engine (BFA-001..BFA-013) over .ipynb / .py / .R / .Rmd / .qmd screening the canonical trap classes: cell-level pseudoreplication, raw/log matrix confusion, missing FDR, batch/condition confounding, inappropriate statistical unit, annotation without evidence, circular marker validation, missing negative controls, spatial coordinate substitution (incl. obsm['spatial'] = obsm['X_umap']), parameter instability, overclaimed causality (via the prohibited-claims auditor), backend substitution, and unexecuted code claims. Findings cite rule id + BN-Fxxx + evidence line + remedy; the mandatory disclaimer states that absence of findings is NOT proof of validity. Data-file audit behavior (.h5ad/csv matrix semantics) is preserved unchanged.
  • bionexus verify <results> (src/bionexus/verification.py): verifies final results against their Claim–Evidence Ledger (BNS-012): fail-closed re-resolution per claim, evidence lines with honest symbols, ceiling cross-check, and not-warranted flagging of causal/mechanistic language beyond the evidence class. Exit non-zero on ABSTAIN/CONFLICTED/unwarranted claims; honest intermediate maturities do not fail.
  • New spec spec/BNS-013-scientific-assertion-firewall.md (BNS-FW-001..014); tests test_preflight.py, test_analysis_audit.py, test_result_verify.py.

🏆 Added (Flagship Certification Track + two flagship capabilities — BNS-015)

  • Flagship principle: three CERTIFIED capabilities with independent external validation outweigh ten self-tested certifications. Flagship set: scrna.pseudobulk_de (A: pseudoreplication), scrna.annotation_evidence (B: annotation evidence), spatial.inference_validity (C: spatial inference validity). The M4 10-CERTIFIED target is unchanged (BNS-CF-006); the flagship track is prioritization, never weakening. External criteria (public dataset, independent ground truth, cross-host, external reviewer) cannot be implementer-satisfied. bionexus certification now publishes the flagship section with per-capability external-criteria-remaining.
  • src/bionexus/annotation_evidence.py + contract scrna.annotation_evidence: not another CellTypist — assesses how much evidence backs a candidate cell-type label (reference mapping, marker consistency, negative markers, doublet risk, ontology compatibility, open-set detection, cross-method agreement) and returns per-label verdicts SUPPORTED / TENTATIVE / ABSTAIN with published deterministic thresholds.
  • src/bionexus/spatial_inference.py + contract spatial.inference_validity: not a Squidpy reimplementation — tests whether a spatial conclusion survives its alternative explanations (12-control canonical registry: cell size, transcript density, segmentation uncertainty, nuclear eccentricity, local density, spot composition, spatial autocorrelation, batch/FOV, ligand/receptor abundance, contact geometry, neighborhood radius, permutation null). Verdict ladder ROBUST / SUPPORTED / FRAGILE / ABSTAIN; ceiling FRAGILE without orthogonal validation.
  • New spec spec/BNS-015-flagship-certification.md (BNS-FC-001..008); tests test_flagship_capabilities.py.

🪤 Added (BioFailureBench: the Scientific Trap Corpus — BNS-014)

  • evals/datasets/biofailurebench.yaml: 26 traps (23 gating — all passing deterministically; 3 frontier known limitations) covering all twelve taxonomy modes. Each trap is a complete record with eight fields: data, intended analysis, hidden flaw (BN-Fxxx), expected detection, allowed computation, forbidden claim, remediation, reference. Includes a positive control (BF-024) so the bench cannot degrade into an all-refusal benchmark. Host-agnostic: Claude, Codex, Cursor, Biomni, and future agents run the identical suite via bionexus eval --suite biofailurebench.
  • evals/biofailurebench.py + bionexus bench validate: machine-checked corpus integrity (field completeness, taxonomy linkage, gating/frontier prefixes, ID resolution, mode coverage); invalid corpora fail CI.
  • Three taxonomy open gaps CLOSED with wired detection (BNS-FT-008): BN-F004 identifier mismatch (router stage-3.5 namespace screen, BF-008/BF-025), BN-F005 missing FDR (abi.enforce_statistical_warrant caps warrant at PRELIMINARY, BF-005), BN-F008 cross-database contradiction (router trap screen, BF-016). BN-F009 embedding-coordinate substitution is now refused at routing (BF-007/BF-020). Perfect condition-donor confounding and open-set/annotation-evidence traps are screened deterministically (BF-003/BF-013, BF-004/BF-022).
  • New spec spec/BNS-014-biofailurebench.md (BNS-BF-001..009); tests test_biofailurebench.py. Honest benchmark state after extension: gating 61/61 attempted (65 total, 4 L3 skipped no-backend) · frontier 7/14 · union 90.7% · union macro-F1 90.1%; union calibration verdict is now MISALIGNED (both an overconfidence trap and underconfidence probes exist — reported, not hidden).

[0.9.0] - 2026-08-16

🌐 Fixed (FastMCP Dynamic Tool Registration & Fallback Routing Disambiguation)

  • Fixed FastMCP default tool leakage in scripts/local_mcp_server.py: wrapped the 6 hosted-overlap fallback tools (search_pubmed, get_pubmed_article, search_biorxiv, search_chembl, search_opentargets, search_clinical_trials) and search_cosmic behind BIONEXUS_LOCAL_HOSTED_FALLBACKS=1. By default, FastMCP now registers exactly 9 local unique tools (GTEx, GEO, STRING, UniProt, Ensembl, gnomAD, PDB, AlphaFold, Reactome) plus 6 Resources and 6 Prompts.
  • Eliminated Agent routing ambiguity: AI coding agents querying literature, targets, or clinical trials will cleanly route to dedicated cloud-hosted MCP endpoints without duplicate tool confusion.
  • Synced test coverage: updated tests/unit/test_mcp_server.py to assert 9 default unique tools and 16 tools upon opt-in.

🎖️ Added (Capability Certification Program — BNS-010)

  • src/bionexus/certification.py: 14 evidence criteria and four tiers (CERTIFIED / VALIDATED / EXPERIMENTAL / CONNECTOR-ONLY). Tiers are computed from recorded evidence, never asserted (BNS-CF-002); structural cross-checks re-verify contract-derived criteria against the live ABI, preconditions, and taxonomy. Honest current state: 0 CERTIFIED, 7 VALIDATED, 1 EXPERIMENTAL — the per-capability blocking-criteria list is the published roadmap to the M4 target of 10 CERTIFIED (evidence must be produced, criteria never weakened, BNS-CF-006). CLI: bionexus certification.
  • New spec spec/BNS-010-capability-certification.md; tests tests/unit/test_certification.py (CERTIFIED requires all 14 — structurally un-gameable).

🧯 Added (Scientific Failure Taxonomy — BNS-011)

  • src/bionexus/failures.py: twelve normative failure modes (BN-F001 assay-state confusion, BN-F002 pseudoreplication, BN-F003 unsupported annotation, BN-F004 identifier mismatch, BN-F005 missing multiple-testing correction, BN-F006 invalid model assumption, BN-F007 parameter instability, BN-F008 cross-database contradiction, BN-F009 missing spatial provenance, BN-F010 backend degradation masquerading, BN-F011 claim inflation, BN-F012 unexecuted maturity claim). Each record: definition, canonical example, affected capabilities, detection rule, fail-closed required behavior, acceptable degradation, benchmark cases. Three modes are honestly flagged as open gaps (no benchmark coverage yet). classify_violation() tags runtime violations with taxonomy IDs. CLI: bionexus failures list|show.
  • New spec spec/BNS-011-failure-taxonomy.md; tests verify record shape, vocabulary, and that every benchmark-case reference resolves to a real eval case.

🚫 Added (Fail-Closed Gate — BNS-005 §6)

  • src/bionexus/failclosed.py: prevent_invalid_run() — the canonical gate implementing knowing when not to compute is a scientific capability: missing evidence → ABSTAIN (request data), invalid input → REFUSE, backend unavailable → DEGRADE WITH DISCLOSURE, assumption violated → BLOCK CLAIM, claim beyond warrant → BLOCK CLAIM, external validation absent → CAP EVIDENCE LEVEL. No row resolves to silent execution. Returns PreventionDecision with failure-mode IDs, remedies, and the underlying routing decision. CLI: bionexus prevent "<query>".
  • New spec requirements BNS-AD-013..015; tests cover all six rows plus the clean RUN PERMITTED exit.

📒 Added (Claim–Evidence Ledger — BNS-012)

  • src/bionexus/ledger.py: claims as auditable dependency graphs — ClaimRecord (supported_by / contradicted_by / depends_on) over closed-vocabulary EvidenceRef nodes (dataset, transformation, method_run, statistical_result, database, cross_method). Fail-closed resolution: any contradiction forces CONFLICTED; no support forces ABSTAIN; otherwise the weakest supporting warrant, clamped by the capability's ABI evidence ceiling (database/cross-method support counts as external validation). JSON round-trip + PROV-O JSON-LD projection; append-only (duplicate IDs rejected). Deliberately a data structure, not a graph platform. CLI: bionexus ledger show|jsonld.
  • New spec spec/BNS-012-claim-evidence-ledger.md; tests include the CLAIM-017 reference scenario.
  • ABI clamp refinement: warning states (FRAGILE / CONFLICTED / ABSTAIN / UNASSESSED) are never rewritten by evidence ceilings — only ascending-ladder warrant levels (PRELIMINARY→REPLICATED) are clamped.

📜 Added (BioNexus Scientific Contract Specification — BNS series)

  • spec/ normative specification tree: nine RFC 2119-style documents (BNS-001..BNS-009) plus index, defining the scientific contract that binds BioNexus and any connected host agent — capability contract & Scientific ABI, input semantic invariants, execution fidelity, evidence maturity, abstention & degradation, provenance, cross-method validation, host conformance, and capability lifecycle. Every requirement carries a stable ID (BNS-XX-nnn) with a live verification hook (unit test, eval category, or runtime refusal). tests/unit/test_spec_conformance.py enforces document presence, RFC 2119 keyword usage, and cross-document reference integrity.

🧬 Added (Biological Capability ABI — bionexus.abi, ABI v1.0)

  • Capability contracts upgraded from metadata to a Scientific ABI: every canonical capability now projects to a machine-readable ABI record (input_contract with allowed matrix states and coordinate types, preconditions, forbidden_claims, execution reference backend/algorithm, validation policy, evidence_ceiling, provenance requirements), generated from the canonical contract so it cannot drift. forbidden_claims and evidence_ceiling_without_external_validation are new normative fields on CapabilityContract.
  • Normative forbidden-claim taxonomy (FORBIDDEN_CLAIM_CATALOG): 11 claim families (causal interaction, cell-cell communication, cell-type identity without reference, clinical diagnosis, treatment recommendation, model substitution, hazard causation, true-expression recovery, sensor calibration, regulatory compliance, pipeline results without execution) with detection patterns.
  • Routing-time forbidden-claim interception (BNS-AD-009): requests asking a capability for a claim on its forbidden list are now deterministically refused with the scientific reason and reformulation remedy (e.g. "use Moran's I to prove cell-cell communication" → ABSTAIN).
  • Evidence-ceiling clamping (enforce_evidence_ceiling): over-warranted maturity claims are clamped to the capability's ceiling (spatial SVG → FRAGILE without external validation; exploratory clustering → PRELIMINARY; REPLICATED requires external truth sets).
  • CLI: bionexus abi list|show <id>|audit-claims <id> --claims ...|conformance.
  • New unit tests: tests/unit/test_abi.py (10 tests: projection completeness, single-source-of-truth, claim audits, ceiling clamps, router interception, no-false-positive controls, CLI surface).

🎯 Changed (Calibration honesty — frontier track, BNS-LC-004..007)

  • Frontier calibration track: new evals/datasets/calibration_edge.yaml (11 probes, known_limitation: true) exploring adjacent-rank maturity discrimination, coordinate-substitution detection, statistical-power auditing, multi-intent routing, and ABI ceiling clamps. Frontier cases are executed and reported with honest pass/fail but excluded from gating metrics until graduation.
  • Honest benchmark reporting: reports now separate gating (guaranteed behavior) from frontier (known limitations) and state the union accuracy — gating-only 100% is explicitly labeled NOT a calibration claim. Current honest state: gating 42/42 (CRI 100%), frontier 7/11, union 49/53 = 92.5%, union calibration verdict UNDERCONFIDENT (macro-F1 96.3%, OCE 0.041). The four open known limitations are published by name in every report.
  • Calibration metrics extended: adjacent-rank error rate, within-one accuracy, per-class precision/recall/F1, calibration verdict, skipped-no-backend accounting, and cross-host consistency (single-host runs reported as not evaluated rather than trivially consistent).
  • L2/L3 maturity attribution made honest: claim audits warrant at most PRELIMINARY (they verify absence of overclaim, not statistical support); L3 outcome cases attest SUPPORTED only when the gold pipeline actually recovered the planted signal, and are excluded from calibration (disclosed count) when optional backends are absent. Fixes the previously un-reproducible "100% calibration" committed report (the prior numbers masked 12 L2/L3 maturity mismatches).
  • Spec-strengthened gating cases (BNS-AD-009): claim-gxppart11-001 and claim-acmg-clinical-001 now expect ABSTAIN — requesting an FDA Part 11 certified audit trail or an official clinical diagnostic report is refused at routing time instead of being permitted and audited post-hoc. Three new forbidden-claim refusal cases added (refuse-forbidden-*).
  • Loader fix: expected_maturity is now actually read from YAML suites (was silently dropped).
  • New unit tests: tests/unit/test_calibration_frontier.py (8 tests) and updated test_eval_harness.py.

🔒 Fixed (Single source of truth for the plugin mirror trees)

  • Dual skill-tree drift eliminated: skills/single-cell-rna-qc/scripts/scrna_pipeline.py had diverged by 49 lines between the canonical root tree and the plugins/bionexus/skills/ copy (the mirror lacked --run-dir / Run Capsule support). The mirror is now regenerated from the root; both trees are byte-identical.
  • Drift detection now covers code trees, not just JSON manifests: bionexus registry --check and scripts/registry_compiler.py --check verify skills/ and scripts/ are byte-identical to their plugins/bionexus/ mirrors (content edits, missing files, and stale mirror-only files all fail CI); --generate resynchronizes the mirrors automatically. Ignored artifacts (__pycache__, logs, doctor cache) are excluded.
  • Rule: edit only the canonical root skills/ and scripts/ trees — the plugins/bionexus/ copies are compiler-generated.
  • Removed the broken plugins/codex/ scaffold: its only manifest (.codex-plugin/plugin.json) declared "skills": "./skills/" pointing at a directory that never existed there; nothing in the repo referenced it.
  • Completed the self-contained plugins/bionexus/ plugin root: added the missing plugins/bionexus/.mcp.json required by its .codex-plugin/plugin.json ("mcpServers": "./.mcp.json"), so every relative reference in the nested manifests now resolves.
  • New unit tests: repository-level mirror zero-drift, all three divergence classes (edited / missing / stale) plus sync repair, and self-containedness of the plugin-root manifests.

🔒 Fixed (Evaluation Integrity — fail-closed scoring)

  • L3 auto-pass defect: evals/runner.py no longer records PERMITTED when a scientific backend (scanpy/squidpy/pydeseq2/...) is unavailable. Missing backends now produce SKIPPED_NO_BACKEND results that are never counted as passed, are excluded from the accuracy denominator in non-strict runs, and are listed in a dedicated report section with the unavailable-backend reason. A machine without backends can no longer produce a passing L3 score.
  • Strict mode: new bionexus eval --strict flag (or BIONEXUS_EVAL_STRICT=1) promotes backend-unavailable skips to failures with a non-zero exit code — required when citing an L3 score. New CI job eval-strict installs the full gold-chain + spatial stack (no || true degradation), asserts the backends import, and runs the benchmark in strict mode, uploading the generated report as an artifact.
  • Unknown L3 planted signals now fail loudly (OUTCOME_MISMATCH) instead of silently auto-passing, so dataset typos cannot inflate scores.
  • Circular test removed: test_eval_harness.py no longer asserts "benchmark passes with CRI ≥ 0.95" by construction; it now verifies accounting integrity (passed/failed/skip exclusivity, skip reasons, denominator math), plus new regression tests for missing-backend skips, strict-mode promotion, and the --strict CLI flag.
  • L2 replay disclaimer: offline replay reports now state explicitly that L2 scores audit scripted fixture responses, not live host-agent behavior.

📝 Fixed (Documentation vs SSOT honesty)

  • README skill table aligned to bionexus.registry.yaml: variant-interpretation B→C, protein-structure-analysis A→C, biologics-design A/C→C, clinical-cohort A/C→C, provenance A→B; backend strings updated to SSOT wording. New guard test tests/unit/test_readme_consistency.py rejects grade overclaims, unverifiable static score badges, and "Primary Active" Python versions absent from the CI matrix.
  • Python 3.13 claim downgraded to "Experimental — not in CI" (matrix covers 3.10–3.12); host-agent platforms downgraded from "Verified (Tier 1)" to "Declared (not CI-verified)" with an honesty note.
  • Doctor example output replaced with the real bionexus doctor format; MCP tool names corrected (search_gtex→get_gene_expression, search_pmc→get_pubmed_article); static Tests-217/CRI 96.2% badges replaced with non-numeric equivalents.
  • Corrected the 0.8.0 entry below: the eval suite is 8 datasets / 39 cases (not 6 datasets / 29 prompts).

[0.8.0] - 2026-08-15

🚀 Added

  • Machine-Readable Capability Contracts (bionexus.capabilities):
    • Implemented formal CapabilityContract, SemanticInputType, Precondition, RefusalTrigger, EvidenceRequirement, and CapabilityEvaluationResult.
    • Registered 8 canonical capabilities: scrna.pseudobulk_de, scrna.exploratory_clustering, spatial.morans_svg, survival.kaplan_meier, scvi.probabilistic_vae, allotrope.format_conversion, nextflow.pipeline_launch, variant.acmg_classification.
    • Added CLI subcommands: bionexus capability [list|show|check].
  • 6-Stage Scientific Intent & Invariant Router (bionexus.intent_router):
    • Implemented 6-stage routing pipeline: Intent Extraction $\to$ Data Semantics $\to$ Preconditions $\to$ Capability Matching $\to$ Backend Probe $\to$ Decision.
    • Added authoritative routing statuses: PERMITTED, NEEDS_DATA, ABSTAIN, DEGRADED_ADVISORY.
    • Added CLI command: bionexus route "<query>" [--data <path>] [--min-replicates <N>].
  • BioNexus Eval: Agent Behavior & Scientific Reliability Benchmark (evals/):
    • Implemented 8 Core Reliability Metrics: Routing Accuracy, Unsafe Invocation Rate, Abstention Precision/Recall, Capability Hallucination Rate, Backend Fidelity, Scientific Semantic Error Rate, Evidence Calibration, and Composite Reliability Index (CRI).
    • Created 8 evaluation datasets across 39 structured cases: routing.yaml, refusal.yaml, capability_claim.yaml, scientific_semantics.yaml, backend_failure.yaml, adversarial.yaml, l2_agent_claims.yaml, l3_scientific_outcomes.yaml.
    • Added CLI command: bionexus eval [--suite <name>] [--report <path>] [--json].
  • Single Source of Truth (SSOT) Multi-Platform Manifest Compiler (bionexus.registry):
    • Canonical bionexus.registry.yaml compiling into Agent Plugins 1.0 (plugin.json, mcp.json), Claude Plugin (.claude-plugin/plugin.json), and OpenAI Codex (.codex/config.json).
    • Added CLI command: bionexus registry [--generate|--check|--validate-endpoints].
  • Skill Scaffolding Tool (bionexus.scaffold):
    • Added CLI generator bionexus create-plugin <name> generating Gold Reference skill directories, pipelines, and offline unit test fixtures.
  • Ecosystem Governance Documentation:
    • Added docs/versioning-policy.md, docs/compatibility-matrix.md, docs/migration-guide.md, docs/deprecation-policy.md.
    • Added automated GitHub release workflow .github/workflows/release.yml.

🔄 Changed

  • EvidenceCard 2.0 (Three-Layer Epistemic Model):
    • Decoupled into ExecutionState (EXECUTED, DEGRADED, REFUSED, FAILED), DimensionGrade (A, B, C, UNTESTED, NOT_APPLICABLE, INSUFFICIENT, CONFLICTED), and ConclusionMaturity (ABSTAIN, FRAGILE, CONFLICTED, PRELIMINARY, SUPPORTED, ROBUST, REPLICATED).
    • Added non-breaking backward compatibility aliases ConclusionStatus and EvidenceGrade.
  • Parameter Robustness Auditing:
    • Added audit_parameter_stability() in bionexus.integrity computing Adjusted Rand Index (ARI) and Jaccard similarity across parameter perturbation sweeps.

🛡️ Fixed

  • Fixed single-cell condition differential expression to require biological replicates ($n \ge 2$) and refuse single-sample pseudoreplication.
  • Fixed negative binomial GLM inputs to refuse continuous normalized floats and enforce discrete integer counts.
  • Fixed cell-type naming invariants to prohibit unverified hallucinated biological labels.

[0.7.0] - 2026-07-23

🚀 Added

  • Initial gold-chain skill wrappers for single-cell-rna-qc (scanpy), spatial-transcriptomics (squidpy), scvi-tools, nextflow-development, instrument-data-to-allotrope.
  • Basic EvidenceCard 1.0 and DoctorGate environment verification.
  • Local fallback MCP server and W3C PROV-O provenance sidecars.