do

name	do
description	Execute a bounded ad-hoc engineering task such as a small fix, refactor, docs update, or example without the full run-plan process.

Codex Runtime Rules

These Z Skills are native Codex workflows derived from github.com/zeveck/zskills at commit 14dea81da487b2904ea7d69a27295f1869206cdf. They are invoked by natural language or by naming the skill; examples that look like slash commands are only shorthand for the skill name.

Use Codex behavior first:

• Read the current repository before acting. Prefer existing project scripts, config, and conventions over the upstream examples.
• Never use Claude-only tools or assumptions: no CronCreate, CronList, CronDelete, Agent/Task tool syntax, allowed-tools, .claude/settings.json, or automatic Claude hooks.
• Use sub-agents only when the user explicitly asks for agents, parallel work, or delegation, or when the user explicitly invokes a Z Skill step whose workflow requires an independent reviewer, devil's-advocate critique, or fresh verification context. Keep those skill-required agents bounded to that workflow step. If agents are unavailable or the work is too tightly coupled, run inline, disclose lower assurance before or in the result, and do not auto-land unless the user accepts it.
• For isolation, create git worktrees explicitly with normal git worktree commands. Do not rely on an isolation: "worktree" parameter.
• Config lookup order is project .agents/zskills-config.json first, then project zskills-config.json, then legacy .codex/zskills-config.json, then legacy .claude/zskills-config.json only if already present. Do not create new .claude runtime config for Codex.
• Scheduling is not automatic in Codex. If the user asks for recurring runs, explain the schedule and ask before installing any local cron/system scheduler. For normal turns, perform the requested work now.
• Helper assets live at project .agents/zskills-support by default. Use project-local scripts/* first; fall back to $CODEX_HOME/zskills-support only for explicit global installs or legacy setups.
• Preserve Codex safety rules: do not revert unrelated work, stage files by name, avoid destructive git commands, and verify from actual diffs/tests.

Detailed upstream text is archived in references/upstream-claude-adapted.md for edge cases and future diffs. Load it only when the concise workflow below is insufficient.

Do

Workflow

1. Parse the user request as the task description plus optional mode words: worktree, push, pr, or now. Ignore every as scheduling unless the user explicitly asks to configure a local scheduler.
2. Decide whether the task is small enough for do. If it needs multiple dependent phases, switch to draft-plan or run-plan and say why.
3. Choose execution mode:
   • Default: work in the current repository.
   • worktree or pr: create a git worktree manually under /tmp/<project>-do-<slug> on a named branch.
   • push: verify before pushing.
4. Implement using normal Codex coding discipline: inspect files, make scoped edits, and avoid unrelated changes.
5. Verify with the smallest meaningful test set, plus lint/typecheck/manual checks when relevant.
6. Commit or push only if requested.

Landing Modes

Use the shared Z Skills config lookup order: .agents/zskills-config.json, zskills-config.json, legacy .codex/zskills-config.json, then legacy .claude/zskills-config.json.

• direct: work in the current tree. Refuse direct main commits when execution.main_protected is true unless the user explicitly overrides after warning.
• cherry-pick: for worktree mode, create a manual worktree, commit there, verify, then cherry-pick the scoped commit back to main.
• pr / locked-main-pr: create a manual worktree branch using execution.branch_prefix, push it, and create a PR with gh if available. Do not push to main.

Scheduling

Codex has no built-in Z Skills cron tools. For every, next, or stop, explain that no Codex schedule exists by default. Configure system cron only after a clear user request.

Agents

Use sub-agents only when the user explicitly asked for agent delegation, when a user-invoked Z Skill workflow requires bounded independent review, or when a fresh verifier is allowed by the current Codex delegation policy for a Z Skills landing gate or explicit verification. Otherwise execute inline and disclose the review mode.

Preserved Z Skills Invariants

• Keep this lighter than run-plan; escalate multi-phase work.
• Parse trailing mode words conservatively.
• push requires verification first.
• PR mode uses an isolated branch/worktree unless the user asks otherwise.