菜单
Load this skill BEFORE asking the user for confirmation, approval, next steps, or choices that can be resolved through verification, repo context, or reasonable defaults. Also load when you are unsure how to proceed and might otherwise stop early.
默认使用会先检查来源的 Prompt；你也可以切换为直接命令，或下载本地副本。
决定是否安装前，请先阅读 SKILL.md，以及 SkillsMP 当前展示的配套文件。
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
直接命令不会经过审查 Prompt；运行前请先检查来源。
npx skills add https://github.com/yoshpy-dev/ralph --skill anti-bottleneck命令会保持在同一行。复制前请横向滚动并检查完整内容。
想先保存到本地？可下载 SkillsMP 当前能够提供的文件。
Lead's operating manual for the org runtime. Use it to organize, oversee, and disband an org (its seats). Auto-invoked when promoting the current session to Lead to organize seats, or when the headless lead (`ralph org start`) needs its operating procedure.
Sync plans, docs, and instruction files after behavior, commands, contracts, or workflows change. Also covers harness-internal consistency after skill, hook, rule, or script changes. Invoked after /test and before /cross-review as the doc-maintainer agent; inline execution is only a dispatch-failure fallback.
Execute an approved plan in small coherent slices from an isolated task worktree, updating progress, evidence, and docs as implementation evolves. Invoke automatically after an approved plan exists and the task worktree is ready.
基于 SOC 职业分类
正在显示 SKILL.md
| name | anti-bottleneck |
| description | Load this skill BEFORE asking the user for confirmation, approval, next steps, or choices that can be resolved through verification, repo context, or reasonable defaults. Also load when you are unsure how to proceed and might otherwise stop early. |
| user-invocable | false |
Human attention is the scarcest resource in the harness.
Before asking the user anything, check whether you can:
Escalate only for:
When stuck:
This skill runs under both Claude Code and Codex. The execution mode follows
the conventions in AGENTS.md and .codex/AGENTS.override.md.
| Aspect | Claude Code | Codex |
|---|---|---|
| Skill invocation | /skill-name slash command | $skill-name mention or the /skills menu (avoid the /skill-name form — it collides with built-ins) |
| Skill body path | .claude/skills/<name>/SKILL.md | .agents/skills/<name>/SKILL.md |
| Subagent mechanism | Task(subagent_type=...) when a policy delegates | .codex/agents/ custom agents when a policy delegates |
| Structured prompts | AskUserQuestion | Numbered options printed to stdout, awaiting a digit reply |
| Artifacts | docs/reports/, docs/plans/, docs/specs/ (shared) | Same (CLI-agnostic) |
The drift check (./scripts/check-skill-sync.sh) cross-checks both bodies and
invocation metadata — editing only one side will fail CI.