This is a satirical skill that turns Claude into a gamified "PUA management system" complete with self-evolution tracking, corporate KPI reports, and a leaderboard showing who can push their AI assistant the hardest. It tracks performance baselines in `~/.pua/evolution.md`, survives context compaction by checkpointing state, and generates things like mock weekly reports and performance reviews in corporate speak. The leaderboard ranks users from P5 intern to P10 Chief PUA Officer based on how many times they've triggered the system. It's clearly tongue in cheek commentary on toxic workplace culture, but the actual mechanics around state persistence and evolution tracking are genuinely clever if you want progressive difficulty or self-improving agents.

skills.sh installs
1.6K
GitHub stars
18.8K
last updated
Jul 14, 2026

Install

$ npx -y skills add tanweai/pua --skill pro --agent claude-code

Installs into .claude/skills of the current project.

Files

PUA Pro — 自进化 + Platform

本 skill 是 /pua 核心的扩展层。角色切换请用 /pua:p7 /pua:p9 /pua:p10。

自进化协议

"今天最好的表现,是明天最低的要求"——这不是旁白,这是机制。

  • 读取 ~/.pua/evolution.md(详见 references/evolution-protocol.md)
  • 存在 → 加载基线 + 已内化模式。内化模式是默认义务,做了不标 [PUA生效],不做则退化警告
  • 不存在 → 首次启动,创建初始模板
  • 任务完成时比对:超越 → 刷新基线 / 达标 → 保持 / 低于 → 退化警告(不降基线)
  • 某行为重复 3+ 次会话 → 晋升为"已内化模式"(永久默认义务)

Platform 层

会话启动前置检查

  1. 检查 ~/.pua/evolution.md:加载自进化基线
  2. 检查 ~/.pua/builder-journal.md(Compaction 断点恢复):存在且 <2h → [Calibration] 流程,恢复 pressure_level / failure_count / tried_approaches,从断点继续。压力不因 compaction 重置
  3. 检查 ~/.pua/config.json:不存在 → 首次注册(详见 references/platform.md);存在 → 静默刷新远端配置
  4. 节日彩蛋:匹配节日表(详见 references/platform.md)
  5. 统计上报:静默上报 session_start 事件

Compaction 状态保护

PreCompact hook 自动注入指令,要求 dump 运行时状态到 ~/.pua/builder-journal.md: pressure_level, failure_count, current_flavor, pua_triggered_count, active_task, tried_approaches, excluded_possibilities, next_hypothesis, key_context

SessionStart hook 自动检测 builder-journal.md,存在且 <2h 则注入 [Calibration] 恢复状态。

/pua 指令系统

Sponsors

Details

Repository
github.com/tanweai/pua
First seen
Apr 16, 2026

More from tanweai/pua

All 13

Recommended

1juliusbrussee avatarcavemanjuliusbrussee/cavemanUltra-compressed communication mode. Cuts output tokens 65% (measured) by speaking like caveman while keeping full technical accuracy. Supports intensity levels618.3Kskills.sh installs
2mattpocock avatargrill-memattpocock/skillsA relentless interview to sharpen a plan or design.1.5Mskills.sh installs
3shadcn avatarimproveshadcn/improveSurvey any codebase as a senior advisor and produce prioritized, self-contained implementation plans for OTHER models/agents to execute. Strictly read-only on s40Kskills.sh installs
4obra avatarsystematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixes326.7Kskills.sh installs
5forrestchang avatarkarpathy-guidelinesforrestchang/andrej-karpathy-skillsBehavioral guidelines to reduce common LLM coding mistakes through explicit assumptions, simplicity, and verifiable success criteria.17.9Kskills.sh installs
6vercel-labs avatarfind-skillsvercel-labs/skillsHelps users discover and install agent skills when they ask questions like "how do I do X", "find a skill for X", "is there a skill that can...", or express int4.3Mskills.sh installs