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
The quality sweep creates a persistent dashboard issue (e.g., superdav42/tgc.church#14) listing ShellCheck findings. Minutes later, the supervisor pulse independently creates an actionable issue (#17) for the same findings and dispatches a worker. The worker fixes the code (PR #18), closes #17, but the dashboard issue #14 retains the stale findings in its body. A human later reads #14, sees the unresolved findings, and manually dispatches another worker — which discovers the work is already done.
Human reads #14, sees stale findings, dispatches duplicate /full-loop
Root cause
Two gaps in the quality sweep → supervisor → worker pipeline:
No cross-reference between dashboard and actionable issues. The supervisor pulse creates actionable issues from its own analysis without checking whether the persistent quality dashboard already tracks the same findings. Neither system deduplicates against the other.
The supervisor should check the quality dashboard issue before creating actionable issues for findings already listed there (or vice versa — the dashboard should link to the actionable issue).
When a fix PR merges, the dashboard issue body should be updated (or a comment added) to mark those findings as resolved.
A human reading the dashboard should be able to distinguish resolved vs. unresolved findings.
Proposed fixes
Dedup at actionable issue creation: Before the supervisor creates an issue for a quality finding, search for existing open issues (including persistent dashboard issues) that reference the same file + ShellCheck code. Skip or cross-reference if found.
Post-merge dashboard update: When a PR closes a quality-related issue, trigger an update to the persistent dashboard issue — either re-run the sweep to refresh findings, or add a comment noting which findings were resolved by which PR.
Staleness indicator: Add a "last verified" timestamp to dashboard findings so readers can tell whether findings predate recent merges.
Description
The quality sweep creates a persistent dashboard issue (e.g.,
superdav42/tgc.church#14) listing ShellCheck findings. Minutes later, the supervisor pulse independently creates an actionable issue (#17) for the same findings and dispatches a worker. The worker fixes the code (PR#18), closes#17, but the dashboard issue#14retains the stale findings in its body. A human later reads#14, sees the unresolved findings, and manually dispatches another worker — which discovers the work is already done.Timeline (superdav42/tgc.church, 2026-03-27)
/full-loopRoot cause
Two gaps in the quality sweep → supervisor → worker pipeline:
No cross-reference between dashboard and actionable issues. The supervisor pulse creates actionable issues from its own analysis without checking whether the persistent quality dashboard already tracks the same findings. Neither system deduplicates against the other.
Dashboard not updated after fixes merge. When PR feat(security): add max-security OpenCode GitHub AI agent integration #18 merged and closed docs: update branch creation to recommend worktrees for parallel sessions #17, the persistent dashboard issue feat: Ralph Loop iterative AI development workflows #14 was not updated to reflect that the SC2181/SC2155 findings are resolved. Its body still shows them as open findings.
Expected behavior
Proposed fixes
Related issues
Environment