Operator feedback queue

▼ items are operator-flagged "this feels off" picks from /decide. ▲ items are explicit-positive votes (most useful on weak-LLM rows where you're overriding model skepticism). Each row carries an immutable decision_context snapshot of what you saw at click time.
Workflow: walk these in a Claude Code session, investigate ▼ items, mark resolution per row.

Resolution:open (0)investigated (0)won't fix (0)fix applied (0)all (0)

No feedback rows under this filter. Click ▲/▼ on /decide to add.