// 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.
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.
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-worktree-intake |
| description | 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. |
Use this skill before non-trivial TSZ work. The goal is to avoid wasted cycles from stale branches, hidden dirty files, unowned PR overlap, missing TypeScript corpus state, and cache-destroying cleanup.
Pick or confirm a stable AgentName.
Refresh the remote truth:
git fetch origin main
Read the durable direction and lane state:
sed -n '1,260p' docs/plan/ROADMAP.md
scripts/agents/show-goal.sh <AgentName>
Check local workspace health:
git status --short --branch
scripts/agents/disk-preflight.sh <AgentName>
git worktree list
If disk is low, cache state is unclear, or a new worktree might be needed,
use tsz-disk-cache-hygiene before cleaning or creating anything.
Check coordination overlap before editing:
scripts/agents/list-owned-work.sh <AgentName>
gh pr list --state open --limit 100 --json number,title,isDraft,headRefName,baseRefName,labels,updatedAt,url
gh issue list --state open --limit 100 --json number,title,labels,updatedAt,url
Read references/common-failure-modes.md when any command reveals a stale
merged branch, dirty workspace, low disk, missing TypeScript corpus, or
overlapping active PR.
Use the current branch only when it is the active PR branch for the requested
work. If the current branch's PR is merged, closed, unrelated, or heavily ahead
of origin/main, create a fresh branch from origin/main.
Prefer:
git switch -c codex/<short-scope>-<yyyymmdd> origin/main
If dirty files exist, inspect them before switching. Carry them only when they belong to the requested work; otherwise leave them unstaged and do not include them in the PR.
Reuse an existing sister worktree when disk or TypeScript corpus state makes it cheaper. Create new worktrees beside the checkout, not inside it:
git worktree add ../tsz-<short-scope> -b codex/<short-scope>-<yyyymmdd> origin/main
cd ../tsz-<short-scope>
scripts/setup/link-ts-submodule.sh
Do not run broad du archaeology or cargo clean as a reflex. If disk is low,
follow the cleanup ladder printed by scripts/agents/disk-preflight.sh.
Before coding, state the next action in one sentence:
For behavior work, also state the structural rule and owning layer before editing. For process or docs work, state the iteration bottleneck and the evidence that it is recurring.
Pause and gather stronger evidence when: