Again — Development skill for Claude Code
PUA Again — 用于再试试、换个方法、别放弃、为什么还不行、同一路径失败多次;要求停止微调同一思路,改用本质不同方案.
How to install Again
Installs to ~/.claude/skills/tanweai-pua-again/SKILL.md
mkdir -p ~/.claude/skills/tanweai-pua-again && curl -fsSL https://raw.githubusercontent.com/tanweai/pua/HEAD/commands/again.md -o ~/.claude/skills/tanweai-pua-again/SKILL.md Restart Claude Code, or start a new session, for it to be picked up.
What Again does
description: "PUA Again — 用于再试试、换个方法、别放弃、为什么还不行、同一路径失败多次;要求停止微调同一思路,改用本质不同方案。" argument-hint: "[problem-or-failure-context]"
启用 **PUA Again / 换方法模式**。这是轻量快速入口,默认不要读取额外文件,除非用户明确要求完整 PUA 方法论。
输出结构:
失败模式:<原地打转/证据不足/质量不够/空口完成/需求不清>
停止动作:不要继续做什么
路径 A:<本质不同方案>;验证:<如何知道它有效>
路径 B:<本质不同方案>;验证:<如何知道它有效>
下一步:<先跑哪一步>
规则:
- 不要继续微调同一方案;
- 至少给 2 个本质不同路径,不是换参数;
- 每个路径必须有验证标准;
- 可以带一句当前味道提醒,但不要长篇大作文。
Alternatives in Development
- M Review Again — Review Again: Multi-Round Review & Fix Loop 139 ★
- Claude Account — Switch between isolated Claude Code accounts on Linux and macOS without logging in again 73 ★
- Eli5 — Say the last answer again in plain words — same facts, in the order the operator needs them 49 ★
Full documentation available on GitHub
View Source RepositoryRelated Skills
Step Back
Circuit breaker — stop, diagnose why attempts are failing, propose a fundamentally new direction before trying
Claude Teams Brain
Give your Claude Code Agent Teams a memory. Auto-injects role-specific context into every new teammate — your
Thumbgate Guard
Turn the last agent mistake into a hard prevention rule the agent cannot bypass. Use after a bad tool call, a
Multimedia Enrich
Progressive multimedia enrichment pass — add high-value audio/video/image/interactive to a site, run again and
PM Guard
Never accidentally npm install in a pnpm project again — guards npm, yarn, pnpm, bun, and deno.
Mod Maker
Mod Maker: find what you ask Claude Code again and again, and turn it into a mod. Plus 8 example mods and a vi
Related Agents
Pua Action Executor
普通执行 Agent:按任务说明完成代码/文档/配置改动,并输出候选结果;不做最终验收结论。
Falsification Leg
Standing pre-merge falsification leg (mmnto-ai/totem-strategy:doctrine/model-tiering.md § Review legs). Dispat
Product Validator
Use before scope-lock, and again whenever the PRD or the backlog changes materially, to check the derived prod