fix: resolve open CodeQL alerts (float widening, workflow permissions)

Two high alerts (cpp/integer-multiplication-cast-to-long): forest serving
multiplied f->lr (float) by a float tree value before widening to the double
accumulator, so the product could overflow to infinity, and under x87 excess
precision vary by platform, before the widening. Promote to double before the
multiply in both spots, matching the trainer (predict-train.c already does).

Two medium alerts (actions/missing-workflow-permissions): wheels.yml had no
permissions block, so the build jobs got the default token scopes. Default the
workflow token to contents:read; the publish job keeps its id-token:write
override for PyPI OIDC.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>

Author: mstrathman

.github/workflows/wheels.yml

@@ -5,6 +5,11 @@ on:
     tags: ["v*"]
   workflow_dispatch:
 
+# Default the GITHUB_TOKEN to read-only; the publish job opts into
+# id-token: write for PyPI OIDC and nothing else needs write.
+permissions:
+  contents: read
+
 # Builds the Python wheels (`pip install sqlite-predict`). Each wheel bundles the
 # zero-dependency loadable compiled for its platform. Staging runs `make
 # python-src` on the host to drop the amalgamation + SQLite ext headers into the

CHANGELOG.md

@@ -6,6 +6,15 @@ project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
 ## [Unreleased]
 
+### Fixed
+
+- **Forest prediction accumulates the learning-rate product in double**, matching
+  the trainer. The serving path computed `float * float` before widening to the
+  `double` accumulator, which could overflow to infinity (and, under x87 excess
+  precision, differ by platform) before the widening. Predictions are computed
+  consistently in double now; served values may shift in their least-significant
+  digits.
+
 ## [0.2.0] - 2026-07-31
 
 ### Changed

predict-student.c

@@ -276,7 +276,10 @@ int predict0_forest_predict_row(const Forest *f, const f32 *x, f64 *scbuf, char
   if (f->task == 1) {
     f64 s = f->init[0];
     for (int j = 0; j < f->n_trees; j++)
-      s += f->lr * forest_tree_value(f, j, x);
+      /* (f64) before the multiply: keep the product in double, matching the
+       * trainer (predict-train.c), so a float*float never overflows or picks up
+       * platform-dependent excess precision before it reaches the accumulator. */
+      s += (f64)f->lr * forest_tree_value(f, j, x);
     *has_conf = 0;
     *pred = sqlite3_mprintf("%.17g", s);
     return *pred ? SQLITE_OK : SQLITE_NOMEM;
@@ -285,7 +288,8 @@ int predict0_forest_predict_row(const Forest *f, const f32 *x, f64 *scbuf, char
     scbuf[c] = f->init[c];
   for (int r = 0; r < f->n_rounds; r++)
     for (int s = 0; s < f->n_score; s++)
-      scbuf[s] += f->lr * forest_tree_value(f, r * f->n_score + s, x);
+      /* (f64) before the multiply, as above and in the trainer */
+      scbuf[s] += (f64)f->lr * forest_tree_value(f, r * f->n_score + s, x);
   f64 mx = scbuf[0];
   int arg = 0;
   for (int c = 1; c < f->n_score; c++)