Stop repeated LHP decisions and queue starvation#41
Conversation
|
Failed to generate code suggestions for PR |
|
@codex review |
|
Failed to generate code suggestions for PR |
There was a problem hiding this comment.
💡 Codex Review
Here are some automated review suggestions for this pull request.
Reviewed commit: 772b6485fb
ℹ️ About Codex in GitHub
Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you
- Open a pull request for review
- Mark a draft as ready
- Comment "@codex review".
If Codex has suggestions, it will comment; otherwise it will react with 👍.
Codex can also answer questions or update the PR. Try commenting "@codex address that feedback".
|
@codex review |
|
Codex Review: Didn't find any major issues. More of your lovely PRs please. Reviewed commit: ℹ️ About Codex in GitHubYour team has set up Codex to review pull requests in this repo. Reviews are triggered when you
If Codex has suggestions, it will comment; otherwise it will react with 👍. Codex can also answer questions or update the PR. Try commenting "@codex address that feedback". |
What changed
loop:approved, addloop:needs-human, and allow the queue to advanceWhy
AS215932/network-operations#458 accumulated 460 repeated bot comments. Mutable LHP case/verification state and rotating knowledge artifacts changed the Governor record identity, local-only dedupe could be lost, and a failed top queue item retained
loop:approved, so subsequent cycles selected it again.Impact
A semantic request produces at most one Governor decision even as telemetry changes. Failed or completed daemon attempts are parked for human review, so they cannot starve later approved work. Disk capacity/configuration changes remain explicitly human-gated.
Validation
uvx ruff check src testsuv run --group dev mypy --strict src/var/tmptest root because host/tmpquota was exhausted)Incident: AS215932/network-operations#458