This sheet is synthetic
The tracker, the issues and the dates are invented. The example shows the columns and the four states. Nothing is closed by the sheet; a person approves a batch.
The sheet
One row per issue. Every state carries its evidence.
If the matrix is wider than the box, scroll horizontally to read every column. Keyboard: focus the matrix and use Left/Right.
SYNTHETIC EXAMPLE: five rows from an invented backlog of 412 issues. Stale rule: 365 days without activity.
ROW ISSUE STATE EVIDENCE LABELS (suggested)
1 #118 keep reproduced in comments by two users in the last 60 d bug, checkout
2 #203 possible duplicate same error text as #187; both show "token expired" (link both)
3 #251 needs information no version or steps; question: "Which version?" question
4 #77 proposed stale last activity 640 days ago; no comments since (none)
5 #302 keep one comment from a maintainer 12 days ago feature
DUPLICATES VIEW
#203 ~ #187 matching text: "Error: token expired at refresh" in both bodies
CLOSURE BATCH (for approval)
#77 proposed stale reopen list: yes
SUMMARY BY STATE (412 issues)
keep 151 possible duplicate 38 needs information 97 proposed stale 126
Nothing was closed or edited. A person on your side approves the batch.Why a person closes
A contributor to a large project proposed treating a stale bot as a one-off cleanup; maintainers in the thread disagreed and the proposal was closed as not planned. The contributor argued that closing by silence alone can discard issues that still matter.
GitHub warns that labels can trigger other automation, so suggested labels stay in the sheet.
- A reopen list undoes a mistake
- Suggested labels stay in the sheet
Sources and limits
- etcd issue 21051: stale bot proposal Checked 2026-10-12.
- A contributor proposed on 25 December 2025 to treat a stale bot as a completed one-off cleanup mission and wrote that silence is not low value.
- Maintainers in the thread disagreed with the proposal, and one said missing details, not neglect, explain stalled issues. The proposal was closed as not planned.
- GitHub Agentic Workflows: AI issue triage example Checked 2026-10-12.
- The example agent holds only contents: read and issues: read, and writes go through safe outputs that check label names and comment content first.
- The page warns that labels can trigger other automation.