diff --git a/extensions/ducksmiles/description.yml b/extensions/ducksmiles/description.yml
index 34e9a6211..7ddd8737a 100644
--- a/extensions/ducksmiles/description.yml
+++ b/extensions/ducksmiles/description.yml
@@ -1,7 +1,7 @@
 extension:
   name: ducksmiles
-  description: Cheminformatics toolkit for DuckDB - SMILES, InChI, MOL/SDF, PDB, SELFIES, Wildman-Crippen LogP, Morgan/ECFP fingerprints, and Tanimoto similarity from SQL
-  version: 0.4.0
+  description: Cheminformatics toolkit for DuckDB - SMILES, InChI, MOL/SDF, PDB, SELFIES; descriptors (LogP, TPSA, QED), Morgan/MACCS fingerprints + similarity, drug-likeness/toxicophore filters, and a native in-silico docking pipeline, from SQL
+  version: 0.5.0
   language: C++
   build: cmake
   license: MIT
@@ -10,7 +10,7 @@ extension:
 
 repo:
   github: nkwork9999/duckSMILES
-  ref: 51c796e9340e70c0ff8592771782916954fb4b08
+  ref: 96c85cc2c895e5fdaecb3e2e30009d1201fa5103
 
 docs:
   hello_world: |
@@ -22,21 +22,32 @@ docs:
     SELECT round(mol_weight('c1ccccc1'), 2);
     -- 78.11
 
-    -- Wildman-Crippen LogP (matches RDKit)
-    SELECT round(logp_crippen('CCO'), 4);
-    -- -0.0014
+    -- Wildman-Crippen LogP and molar refractivity (match RDKit)
+    SELECT round(logp_crippen('CCO'), 4), round(mol_mr('c1ccccc1'), 3);
+    -- -0.0014, 26.442
 
-    -- Morgan/ECFP fingerprint as 2048-bit BLOB (ECFP4 default)
-    SELECT bit_count(CAST(morgan_fp_bits('CC(=O)Oc1ccccc1C(=O)O') AS BIT));
-    -- 26   (aspirin popcount)
+    -- QED drug-likeness score (RDKit Chem.QED, weights_mean)
+    SELECT round(qed('CC(=O)Oc1ccccc1C(=O)O'), 3);
+    -- 0.550   (aspirin)
 
-    -- Tanimoto similarity between two fingerprint BLOBs (no CAST AS BIT needed)
+    -- Drug-likeness rule panels + toxicophore alerts
+    SELECT lipinski_violations('CC(=O)Oc1ccccc1C(=O)O'),
+           druglikeness_pass('CC(=O)Oc1ccccc1C(=O)O', 'veber'),
+           structural_alerts_json('O=Cc1ccccc1');
+    -- 0, 1, ["aldehyde","aldehyde_any"]
+
+    -- Morgan/ECFP fingerprint + similarity between two BLOBs
     SELECT round(tanimoto_bit(morgan_fp_bits('CCO'), morgan_fp_bits('CCN')), 4);
     -- 0.3333
 
-    -- Validate and extract InChI layers
-    SELECT inchi_formula('InChI=1S/C2H4O2/c1-2(3)4/h1H3,(H,3,4)');
-    -- C2H4O2
+    -- In-silico docking: SMILES ligand vs PDB receptor -> ranked poses (JSON)
+    SELECT dock('CCO', :receptor_pdb, 1.7,0.2,0.5, 8,8,8, 20, 42);
+
+    -- Retrospective virtual-screening enrichment over a docked, labelled library
+    SELECT roc_auc(list(score), list(is_active)),
+           enrichment_factor(list(score), list(is_active), 0.01),
+           bedroc(list(score), list(is_active), 20.0)
+    FROM docked_library;
 
     -- Convert SMILES to SELFIES (ML-friendly notation)
     SELECT smiles_to_selfies('CCO');
@@ -47,32 +58,52 @@ docs:
     library dependencies (no RDKit required).
 
     **Supported Formats:**
-    - SMILES: Molecular validation, formula, weight, atom/bond counts, LogP
-    - InChI/InChIKey: Layer extraction, stereochemistry detection, skeleton matching
-    - MOL/SDF: V2000/V3000 block parsing, molecule counting
-    - PDB/CIF/XYZ: Protein structure analysis (atom, chain, residue, model counts)
-    - SELFIES: Bidirectional SMILES-SELFIES conversion for ML pipelines
-
-    **39 scalar SQL functions** for molecular property extraction, format conversion,
-    structural comparison, and physicochemical property prediction. Ideal for
-    cheminformatics datasets, drug discovery pipelines, and molecular ML feature
-    engineering.
-
-    **LogP:** `logp_crippen()` implements the Wildman-Crippen atom-contribution
-    method (110 SMARTS patterns, 68 atom types) and matches RDKit's
-    `Crippen.MolLogP` exactly for small molecules.
-
-    **Morgan / ECFP fingerprint:** `morgan_fp_bits()` ports RDKit's MorganGenerator
-    (layered BFS + hash_combine + dead-atom dedup) to Rust and returns a fixed-width
-    bit vector as `BLOB`. Defaults to ECFP4 (radius=2, 2048 bit); 3-arg overload
-    `morgan_fp_bits(smi, radius, n_bits)` exposes full control.
-
-    **Tanimoto similarity:** `tanimoto_bit(BLOB, BLOB) -> DOUBLE` computes
-    `popcount(a & b) / popcount(a | b)` directly on raw BLOB bytes (no
-    `CAST AS BIT` round-trip), processing 8 bytes at a time via `count_ones()`
-    so it lowers to POPCNT on x86_64 / CNT on aarch64. Mismatched lengths raise
-    a clear `InvalidInputException`; empty-vs-empty returns `0.0` (RDKit
-    convention). The SQL-level `bit_count(a & b)::DOUBLE / bit_count(a | b)`
-    is still available and produces bit-exact identical results.
-
-    **Architecture:** Rust (core logic, 5 crates) + C++ (DuckDB integration via FFI)
+    - SMILES: validation, formula, weight, atom/bond counts, canonicalization, aromaticity
+    - InChI/InChIKey: layer extraction, stereochemistry detection, skeleton matching
+    - MOL/SDF: V2000/V3000 block parsing, 3D geometry, molecule counting
+    - PDB/CIF/XYZ: protein structure analysis (atom, chain, residue, model counts)
+    - SELFIES: bidirectional SMILES-SELFIES conversion for ML pipelines
+
+    **130+ scalar SQL functions** for molecular property extraction, format conversion,
+    structural comparison, fingerprinting, similarity search, drug-likeness /
+    toxicophore filtering, and structure-based virtual screening.
+
+    **Descriptors (RDKit-faithful ports):**
+    - `logp_crippen()` / `mol_mr()` - Wildman-Crippen LogP and molar refractivity.
+    - `tpsa()` - Ertl topological polar surface area.
+    - `qed()` - QED drug-likeness (Bickerton 2012), a port of RDKit `Chem.QED`.
+    - Lipinski set: `num_h_donors()`, `num_h_acceptors()`, `num_rotatable_bonds()`,
+      `fraction_csp3()`, `num_heteroatoms()`, and the full ring-count descriptor family.
+
+    **Fingerprints & similarity:**
+    - `morgan_fp_bits()` (RDKit MorganGenerator, ECFP4 / 2048-bit default) and
+      `maccs_keys()` (167-bit MACCS).
+    - Full bit-fingerprint similarity family over raw BLOBs, faithful to RDKit
+      `DataStructs/BitOps.cpp`: `tanimoto_bit`, `dice_bit`, `cosine_bit`, `kulczynski_bit`,
+      `sokal_bit`, `mcconnaughey_bit`, `asymmetric_bit`, `braun_blanquet_bit`,
+      `russel_bit`, and asymmetric `tversky_bit(a, b, alpha, beta)`.
+
+    **Structure tools:** SMARTS substructure search/counting, Bemis-Murcko and generic
+    scaffolds, scaffold networks, ring-system extraction, maximum common substructure
+    (MCS), and MolHash standardization.
+
+    **Drug-likeness & toxicophore filtering:**
+    - `admet_json()` - a full report: descriptors plus six drug-likeness rule panels.
+    - `druglikeness_pass(smiles, rule)` for Lipinski, Veber, Ghose, Egan, Muegge and
+      Lead-likeness; `lipinski_violations()` for the Rule-of-Five count.
+    - `structural_alerts_json()` / `structural_alert_count()` over a curated
+      Brenk / PAINS reactive-group toxicophore catalogue.
+
+    **In-silico docking pipeline (structure-based virtual screening, end-to-end in SQL):**
+    - `smiles_to_pdbqt()` - 3D conformer generation (distance geometry + an L-BFGS-
+      minimised lite force field with aromatic-ring planarity) emitted as PDBQT.
+    - `prepare_receptor(pdb, ph)` - pH-dependent protonation states + polar-hydrogen
+      addition, so the receptor can donate H-bonds.
+    - `dock(smiles, pdb, cx, cy, cz, sx, sy, sz, n_runs, seed [, ph])` - flexible
+      torsion-tree docking with an AutoDock-Vina-style scoring function and 3D affinity
+      maps; returns ranked poses with coordinates as JSON.
+    - Retrospective-validation harness: `roc_auc`, `enrichment_factor(fraction)`,
+      `bedroc(alpha)` to measure how well a docked library enriches known actives.
+
+    **Architecture:** Rust (core logic, zero external chemistry crates) + C++ (DuckDB
+    integration via FFI).