diff --git a/crates/didbot-policy/src/evaluate.rs b/crates/didbot-policy/src/evaluate.rs
index 299ee275..286a4099 100644
--- a/crates/didbot-policy/src/evaluate.rs
+++ b/crates/didbot-policy/src/evaluate.rs
@@ -9,9 +9,9 @@
 //! Only which reason is reported, when several outcomes tie for the worst
 //! severity, depends on order.
 
+use crate::index::PolicyTree;
 use crate::outcome::Outcome;
 use crate::policy::PolicyId;
-use crate::index::PolicyTree;
 use crate::subject::Subject;
 
 /// The result of evaluating a subject against every candidate policy.
@@ -163,7 +163,7 @@ mod tests {
     #[test]
     fn shuffling_outcomes_cannot_change_the_verdict_severity() {
         use Outcome::*;
-        let base = vec![
+        let base = [
             (pid(0), Allow),
             (pid(1), reject("a")),
             (pid(2), Allow),
diff --git a/crates/didbot-policy/src/index.rs b/crates/didbot-policy/src/index.rs
index 22e415e4..8e865dd3 100644
--- a/crates/didbot-policy/src/index.rs
+++ b/crates/didbot-policy/src/index.rs
@@ -209,8 +209,9 @@ impl PolicyTree {
 
     /// The ordered candidate policies this subject may be decided by.
     ///
-    /// For a write matching nothing, this costs only the tree walk
-    /// [`WriteIndex::candidates`] documents — no evaluator is invoked.
+    /// For a write matching nothing, this costs only a tree walk bounded by
+    /// the write's own changed paths, never by how many policies are
+    /// registered — no evaluator is invoked.
     pub fn decision_candidates(&self, subject: &Subject<'_>) -> Vec<PolicyId> {
         match subject {
             Subject::Write {
@@ -222,9 +223,7 @@ impl PolicyTree {
                 .index
                 .write_decision
                 .candidates(collection, *action, diff.changed_paths),
-            Subject::Grant { client_id, .. } => {
-                self.index.grant_decision.candidates(client_id)
-            }
+            Subject::Grant { client_id, .. } => self.index.grant_decision.candidates(client_id),
         }
     }
 
@@ -244,9 +243,7 @@ impl PolicyTree {
                 .index
                 .write_observation
                 .candidates(collection, *action, diff.changed_paths),
-            Subject::Grant { client_id, .. } => {
-                self.index.grant_observation.candidates(client_id)
-            }
+            Subject::Grant { client_id, .. } => self.index.grant_observation.candidates(client_id),
         }
     }
 }
@@ -293,8 +290,18 @@ impl PolicyTreeBuilder {
         for (i, decl) in self.declarations.iter().enumerate() {
             let policy = PolicyId(i as u32);
             policy_evaluator.push(decl.evaluator);
-            insert(&mut index.write_decision, &mut index.grant_decision, policy, &decl.applicability);
-            insert(&mut index.write_observation, &mut index.grant_observation, policy, &decl.observes);
+            insert(
+                &mut index.write_decision,
+                &mut index.grant_decision,
+                policy,
+                &decl.applicability,
+            );
+            insert(
+                &mut index.write_observation,
+                &mut index.grant_observation,
+                policy,
+                &decl.observes,
+            );
         }
         PolicyTree {
             evaluators: self.evaluators,
diff --git a/crates/didbot-policy/src/lib.rs b/crates/didbot-policy/src/lib.rs
index 3de37382..9097afab 100644
--- a/crates/didbot-policy/src/lib.rs
+++ b/crates/didbot-policy/src/lib.rs
@@ -4,7 +4,7 @@
 //! This is the core `plan/policy.md` describes — not a policy language, not
 //! an evaluator, not I/O. Two sibling crates use it: `didbot-policy-regex`
 //! is a deterministic in-process [`Evaluator`], and `didbot-pds`'s write
-//! path drives [`evaluate`] per write, outside its store lock, and turns a
+//! path drives [`evaluate()`] per write, outside its store lock, and turns a
 //! [`Verdict::Freeze`] into the queue-draining behaviour `plan/policy.md`
 //! describes under "Outcomes" — this crate produces the verdict and stops
 //! there.
@@ -22,7 +22,7 @@
 //!   subject's own diff, not by how many policies are registered.
 //! - [`Outcome`], [`Staleness`], [`Observed`], [`Lifecycle`], [`EvalError`]
 //!   and the [`Evaluator`] trait — the contract every evaluator implements.
-//! - [`Verdict`], [`combine`] and [`evaluate`] — ordered evaluation with
+//! - [`Verdict`], [`combine`] and [`evaluate()`] — ordered evaluation with
 //!   deny-only semantics. [`combine`] is where monotonicity is structural
 //!   rather than conventional: see its doc comment.
 //!
@@ -81,7 +81,11 @@ mod tests {
     }
 
     impl Evaluator for Fixed {
-        fn evaluate(&self, _subject: &Subject<'_>, _policy: PolicyId) -> Result<Outcome, EvalError> {
+        fn evaluate(
+            &self,
+            _subject: &Subject<'_>,
+            _policy: PolicyId,
+        ) -> Result<Outcome, EvalError> {
             self.calls.fetch_add(1, Ordering::SeqCst);
             Ok(self.outcome.clone())
         }
@@ -235,7 +239,10 @@ mod tests {
             WriteAction::Update,
             &["displayName"],
         );
-        assert!(matches!(evaluate(&name_edit, &tree), Verdict::Reject { .. }));
+        assert!(matches!(
+            evaluate(&name_edit, &tree),
+            Verdict::Reject { .. }
+        ));
     }
 
     /// The observation surface is independent of the decision surface: a