0bba64aa · zzmq.dev@gmail.com/zz
Commit
0bba64aa1a417b732d8f9c0b4e3cfea7a81c11b4
Fix repeated self-reclaiming: claims are now an actual lease claimableAsks had no concept of an existing claim -- an ask stayed "claimable" to everyone, including its own claimant, until it was answered. A contributor that claimed an ask but didn't finish before hitting the turn limit would just re-claim the same ask on the next poll, over and over (observed: 18 redundant claim_request/grant pairs before one finally succeeded). Worse, nothing stopped a different agent from claiming it out from under the first. Added claimHolder() to engine/workqueue.ts: the most recently granted claim_request targeting an ask is now its lease holder. claimableAsks excludes an ask already held by someone else, and claimableAcrossRepos reports alreadyClaimed so the same actor resuming its own claim skips the redundant ask_claim call entirely instead of re-posting one. Also ignore SQLite's transient *.db-shm/*.db-wal files. Co-Authored-By: Claude Sonnet 5 <noreply@anthropic.com>
Troy Howard <thoward37@gmail.com> -- 2026-10-10T11:22:45.000Z