A run that runs out of room now fails clearly
An agent run that outgrew its available working space used to be marked complete with nothing done; it now fails with a clear explanation instead.
Summary
When a task was too large for a single agent run to hold in its working memory, the run used to be marked complete even though nothing had actually been done. It now fails with a plain explanation of what happened and what to do next.
What You Gain
Before this fix, a run that ran out of room would quietly report "finished," which looked exactly like success — so you could believe a task was done when it wasn't, sometimes for a while before noticing. Now you see immediately when a task was too big for one run, with a clear next step: split the work into sub-tasks, or start a fresh run.
How It Works
These runs are not retried automatically, since simply running the same request again hits the same limit, and no pull request is opened for a run that produced nothing. The failure message tells you directly what went wrong, so you can decide how to re-scope the work.
Where to Find It
This applies automatically to every agent run — no setting to change.
See what else shipped this month on the Effigate Command Center changelog, or try Effigate Command Center with your own team.