// Drive TSZ GitHub PRs through review, CI, and the merge queue. Use when checking PR status, debugging failing GitHub Actions, deciding whether a PR is WIP, marking draft PRs ready, enqueueing or verifying the queue, landing a PR, or interpreting TSZ CI gates.
Create, refresh, and publish TSZ pull requests with correct coordination state. Use when opening a draft PR, updating a PR body, applying canonical agent labels, changing WIP/draft/ready state, acknowledging review, handing off readiness to a manager agent, or making sure a TSZ PR satisfies AgentName, Project Corpus Impact, verification, and overlap rules.
Prevent TSZ machines and worktrees from running out of disk while preserving useful build caches. Use before creating worktrees, running large builds/tests, diagnosing low disk, cleaning artifacts, deciding whether to delete caches, or recovering space without slowing future Rust/TypeScript builds.
Audit TSZ repo iteration speed and quality. Use when asked to deep-dive repo health, identify avoidable delays, improve agent workflows, reduce CI/PR churn, add guardrails, choose skills to install or create, or turn recurring TSZ process failures into focused tooling/docs PRs.
Use when planning, implementing, reviewing, or interpreting TSZ performance work, including benchmark regressions, cache/residency changes, timing claims, perf counters, hotspot investigations, OOM/timeout/stack-overflow blockers, or optimization PR evidence.
Start TSZ work safely in a crowded multi-agent checkout. Use when beginning a TSZ task, resuming a goal, switching or creating branches/worktrees, recovering from a merged or stale branch, checking disk/cache state, reading lane goals, or deciding whether local dirty files belong to the requested work.
Preserve TSZ architecture boundaries while changing checker, solver, binder, emitter, LSP, WASM, or CLI code. Use when planning semantic fixes, moving logic across query boundaries, handling architecture guard failures, ratcheting boundary debt, or reviewing changes for solver-first and no hardcoding rules.
| name | tsz-ci-pr |
| description | Drive TSZ GitHub PRs through review, CI, and the merge queue. Use when checking PR status, debugging failing GitHub Actions, deciding whether a PR is WIP, marking draft PRs ready, enqueueing or verifying the queue, landing a PR, or interpreting TSZ CI gates. |
Use this skill for TSZ PR operations. The repository uses draft PRs for
coordination and heavy CI only after ready-for-review. TSZ's required
Queue Tested status is produced by the repo-local merge queue after a ready
PR is labeled merge-queue. A queued merge is not a landed merge; always
verify the final PR state.
gh for GitHub operations.AgentName in PR bodies and substantive comments.WIP label, starts with [WIP], or
says it is WIP in the title/body/branch description.[codex] to PR titles.Queue Tested is queue-owned. It may be missing or pending before enqueue;
do not wait for it before labeling an otherwise ready PR merge-queue.state: MERGED or keep working.Inspect merge readiness:
gh pr view <pr> --json state,isDraft,mergeStateStatus,mergeable,autoMergeRequest,mergedAt,labels,title,url,headRefName,headRefOid
gh pr checks <pr> --json name,state,bucket,link,completedAt
Watch checks only when the result is immediately needed:
gh pr checks <pr> --watch --interval 20
Inspect failed logs:
gh run view <run-id> --log-failed
gh run view <run-id> --job <job-id> --log
Confirm the PR is not WIP by state, labels, title, body, and branch context.
Confirm the head SHA matches the checks being inspected.
If checks fail, fix the root cause, push, and comment with what changed.
When implementation and verification are complete, mark ready if still draft:
gh pr ready <pr>
Once PR-head checks such as CI Summary and GitGuardian Security Checks
pass for the exact head, enqueue the PR with the durable queue label:
gh pr edit <pr> --add-label merge-queue
Watch Queue Tested only as queue evidence, not as a PR-head prerequisite.
The Poor Man's Merge Queue workflow synthetic-tests one latest-main
merge branch and posts Queue Tested to the PR head before merging.
After the queue runs, verify the merge actually happened:
gh pr view <pr> --json state,mergedAt,mergedBy,url
If a merge-queue labeled PR remains open after green PR-head checks, inspect
Queue Tested, the Poor Man's Merge Queue workflow, and any synthetic
automation/merge-queue/pr-<n> CI run. Do not direct-merge around the queue
unless the user explicitly asks for an emergency admin bypass.
CI Summary is the required umbrella PR-head status; inspect its
prerequisites.Queue Tested is the required queue status; inspect the queue workflow and
synthetic merge branch when it fails or stays pending.conformance-aggregate can fail even when shards pass if accepted-regression
drift exists. Compare unlisted vs resolved accepted paths.unit-cloudbuild often exposes memory, linking, or architecture-contract
failures that local fast tests may miss.Use concise comments:
AgentName: <name>