You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
🐛 review-fix: follow-up commits pushed by hive/relay agents lack Signed-off-by, bouncing the PR to no-dco / 072 #14605
Follow-up commits pushed by hive agents during the review-fix cycle lack the Signed-off-by trailer, so the PR immediately regresses to no-dco / stage/072-returned-to-contributor (or 070-fixing with reason dco-missing) and prow posts the DCO nag.
Observed today on hosted-available-oke-11-placeholder-r05x (build 6977027a):
In each case the hive then nudges the contributor to git rebase --signoff — for #14596 it is nudging itself. The 072 dwell is 24h, so these look healthy on the dashboard while they are actually waiting on a trailer.
Expected
The review-fix kick prompt (BuildFixPrompt) and the contributor relay fix path should require git commit -s / git rebase --signoff on every follow-up commit, and the gh-wrapper / push guard should refuse to push a hive-owned head whose new commits lack the trailer (a rebase --signoff before push is cheap).
Reusing the DCO check result: when a hive-owned head goes no-dco, route straight back into a bounded fix attempt ("add sign-off") rather than returning it to a contributor that is the hive itself.
🔒 Claimed by agent scanner (agent on hive hosted-available-oke-11-placeholder-r05x). The claim lasts 2h0m0s without activity and renews while the work continues. Reply /unclaim to release, or /claim to take it over if you outrank the holder.
Direction for the fix (PR #14628 is on the right track):
Prevention first.BuildFixPrompt and the relay fix path must instruct git commit -s / git rebase --signoff on every follow-up commit, and the push guard must refuse to push a hive-owned head whose new commits lack Signed-off-by — do a rebase --signoff before push instead of failing.
Self-heal, don't bounce. When a hive-owned or relay-authored head goes no-dco, route into one bounded "add sign-off" fix attempt (070) rather than 072 returned-to-contributor. 072 is for humans; the hive nudging itself is a bug.
Keep the 072 path for genuinely human-authored unsigned commits; don't rewrite anyone else's history.
Add a regression test: fix commit without trailer → guard rewrites/refuses; hive-owned no-dco → 070 not 072.
No config knob needed; this is always-on behaviour.
@clubanderson this issue was closed by @hivecommons-hive[bot] as completed.
If this was closed by a merged PR, the fix only takes effect once it ships in a tagged release and your deployment updates to it — check your running version against the fix before concluding the fix failed.
If you are still seeing the problem, comment /reopen and it will be reopened for you — you do not need write access. Please add what you are still observing (running version, logs, steps) so it is not closed the same way again.
Activity
hivecommons-hive commentedon Oct 27, 2026
🔒 Claimed by agent
scanner(agent on hive hosted-available-oke-11-placeholder-r05x). The claim lasts 2h0m0s without activity and renews while the work continues. Reply/unclaimto release, or/claimto take it over if you outrank the holder.clubanderson commentedon Oct 27, 2026
/hive decision approve
Direction for the fix (PR #14628 is on the right track):
BuildFixPromptand the relay fix path must instructgit commit -s/git rebase --signoffon every follow-up commit, and the push guard must refuse to push a hive-owned head whose new commits lackSigned-off-by— do arebase --signoffbefore push instead of failing.no-dco, route into one bounded "add sign-off" fix attempt (070) rather than 072 returned-to-contributor. 072 is for humans; the hive nudging itself is a bug.No config knob needed; this is always-on behaviour.
github-actions commentedon Oct 28, 2026
@clubanderson this issue was closed by @hivecommons-hive[bot] as completed.
If this was closed by a merged PR, the fix only takes effect once it ships in a tagged release and your deployment updates to it — check your running version against the fix before concluding the fix failed.
If you are still seeing the problem, comment
/reopenand it will be reopened for you — you do not need write access. Please add what you are still observing (running version, logs, steps) so it is not closed the same way again.