Release v0.4.4 · highflame-ai/codeoid

What's Changed

feat(web): one cross-session "Needs you" queue by @saucam in #330
feat(web): show repo, branch and target on a dispatch approval by @saucam in #333
feat(web): per-backend spend on the fleet rail by @saucam in #334
fix: pack-scoped skill ids, dangling-link repair, session-scoped skills via SDK plugins by @saucam in #336
feat: the findings loop — review → fix → re-review in the engine, roles untouched by @saucam in #337
fix: let the orchestrator read the artifacts it is told to synthesize by @saucam in #340
fix: bump the agent SDK to 0.3.281, and unstick Opus from a superseded release by @saucam in #341
feat: make the resume session cap configurable, and raise it to 200 by @saucam in #343
feat: take the context window from the backend instead of guessing it by @saucam in #342
feat(web): a dispatch timeline lens on the fleet rail by @saucam in #335
feat: make the default backend configurable with session.defaultProvider by @saucam in #344
fix: background agents stall the session — adopt the CLI's own turns and let background work outlive the turn by @saucam in #345
fix: decode HTML entities in skill-stderr so an approved grant matches by @WeiYiAcc in #347
feat: OAuth sign-in for remote MCP servers, run by codeoid itself by @saucam in #350
fix(qwen): make Stop actually stop instead of replaying the prompt by @saucam in #351

Full Changelog: v0.4.3...v0.4.4