perf(skills): coalesce inventory before run cache locking (#16576)

Share each active company skill-inventory refresh before taking the version-cache lock. Read actor visibility under the lock and keep snapshot publication protected through durable run pinning. Preserve empty-selection behavior and the list API.

Validation: 176 affected cases, full repository typecheck/build, fresh full CI/E2E, security and exact-head Apex 5/5 pass. The local six-reader latency experiment improves from 50.4s to 18.3s; live response latency is a separate acceptance check.

Co-Authored-By: Paperclip <noreply@paperclip.ing>

doc/DEVELOPING.md

@@ -1506,6 +1506,12 @@ as they are available, so later setup failures still clean up acquired resources
 Busy-account deferral returns explicitly before session/workspace preparation.
 Public-path tests are in `server/src/__tests__/heartbeat-execution-configuration.test.ts`.
 
+Concurrent skill preparations join the company inventory refresh before taking
+the version-cache lock. Each preparation then reads its actor's current skill
+visibility under that lock. Snapshot materialization and the durable run version
+pins stay in the same locked callback. Only the in-flight inventory work is
+shared; completed refreshes and permission results are not cached.
+
 Browser Use skill assignment first checks for an enabled, active Browser Use
 connection in the run's company. With none, it skips full tool discovery.
 With one, the existing task, tool-policy and credential-grant checks still apply.