feat: integrate HNSW into agent search, fix Python 3.14 build

Resolves two gaps found in a project-state review:

1. Python build was broken: PyO3/numpy 0.23 caps at Python 3.13 but the
   environment has 3.14. Bumped to 0.28 and updated the two breaking APIs
   (PyObject -> Py<PyAny>, allow_threads -> detach). The extension module now
   imports and round-trips under Python 3.14, unblocking cargo build --workspace.

2. The "HNSW vector search over agent memories" headline was unwired:
   clawhdf5-ann had zero dependents and the agent used a linear cosine+BM25 scan.
   - clawhdf5-ann is now a live index: insert, mark_deleted (soft delete with a
     deleted bitset, traversed but never returned), compact, and a format
     version tag (v2) with backward-compatible load of v1 files.
   - clawhdf5-agent wires HNSW behind the `hnsw` feature (ON by default). The
     index mirrors the cache (node id == cache index) and self-heals: it rebuilds
     whenever hnsw_synced_len drifts from cache.len(), so unhooked pushes can't
     desync it. Non-indexable stores (no/zero-dim/mixed embeddings) and queries
     whose dim doesn't match fall back to the exact linear scan.
   - hybrid.rs gains merge_vector_keyword, shared by the linear and HNSW paths.
   - tests/hnsw_integration.rs validates recall vs a brute-force oracle plus
     insert/delete/batch behaviour.

Disable HNSW for exact search with `--no-default-features --features float16`.

Co-Authored-By: Claude Opus 4.8 (1M context) <[email protected]>

crates/clawhdf5-agent/src/hybrid.rs

@@ -60,6 +60,22 @@ pub fn hybrid_search(
     };
     let kw_scores = bm25_index.search(query_text, vectors.len());
 
+    merge_vector_keyword(vec_scores, kw_scores, vector_weight, keyword_weight, k)
+}
+
+/// Merge pre-computed vector-similarity and keyword scores into a single ranking.
+///
+/// Both score sets are independently min-max normalized to [0, 1] and combined
+/// with the given weights. This is the shared core of [`hybrid_search`]; it is
+/// also used by the optional HNSW path, which supplies vector scores from an
+/// approximate-nearest-neighbour index instead of a full linear scan.
+pub fn merge_vector_keyword(
+    vec_scores: Vec<(usize, f32)>,
+    kw_scores: Vec<(usize, f32)>,
+    vector_weight: f32,
+    keyword_weight: f32,
+    k: usize,
+) -> Vec<(usize, f32)> {
     // Normalize each set to [0, 1].
     let vec_normalized = normalize_scores(&vec_scores);
     let kw_normalized = normalize_scores(&kw_scores);