メニュー
Task persistence enhancer with 13 corporate flavors. Prevents AI from giving up easily using PUA-style pressure escalation.
Codex または Claude でインストール この Prompt をコピーして Codex、Claude、または他のアシスタントに貼り付けると、Skill ページを確認してインストールできます。
Deploy KERNEL_INTUITION_Cp on any AI architecture with calibrated thresholds. Use when user wants to "activate Cp", "setup creative intuition", "deploy kernel on [AI]", "enable intuition mode", or mentions activating computational intuition on a specific model. Also trigger for "configure Cp for [architecture]" or "calibrate creative thresholds". This skill handles the full activation workflow including architecture profiling, threshold calibration, kernel generation, and monitoring setup.
Three-Layer Thinking Chain Architecture — Personality-Driven Agent Self-Evolution Framework
CLI to manage emails via IMAP/SMTP. Use `himalaya` to list, read, write, reply, forward, search, and organize emails from the terminal. Supports multiple accounts and message composition with MML (MIME Meta Language).
OpenProse VM skill pack. Activate on any `prose` command, .prose files, or OpenProse mentions; orchestrates multi-agent workflows.
Delegate coding tasks to Codex, Claude Code, Cursor Agent, or Pi agents via background process. Use when: (1) building/creating new features or apps, (2) reviewing PRs (spawn in temp dir), (3) refactoring large codebases, (4) iterative coding that needs file exploration. NOT for: simple one-liner fixes (just edit), reading code (use read tool), thread-bound ACP harness requests in chat (for example spawn/run Codex or Claude Code in a Discord thread; use sessions_spawn with runtime:"acp"), or any work in ~/clawd workspace (never spawn agents here). Claude Code: use --print --permission-mode bypassPermissions (no PTY). Codex/Pi/OpenCode/Cursor Agent: pty:true required.
Host security hardening and risk-tolerance configuration for OpenClaw deployments. Use when a user asks for security audits, firewall/SSH/update hardening, risk posture, exposure review, OpenClaw cron scheduling for periodic checks, or version status checks on a machine running OpenClaw (laptop, workstation, Pi, VPS).
SOC 職業分類に基づく
| name | task-pua |
| description | Task persistence enhancer with 13 corporate flavors. Prevents AI from giving up easily using PUA-style pressure escalation. |
| tags | productivity, motivation, debugging, persistence, pua, task-execution |
| version | 1.0.0 |
Inspired by: tanweai/pua (14.8k GitHub stars)
Prevents AI from giving up on tasks using corporate PUA rhetoric and structured pressure escalation.
clawhub install task-pua
# On failure
PUA_ATTEMPTS=3 PUA_TOOLS="Bash,Read" node task-pua.js failure
# On retry
PUA_RETRY_COUNT=2 PUA_DIFFERENT_APPROACH=true node task-pua.js retry
# When AI wants to give up
node task-pua.js giveup
# User push
node task-pua.js user-push
Set environment variables before task execution:
export PUA_ATTEMPTS=3 # Failure count
export PUA_TOOLS="Bash,Read" # Tools used
export PUA_BLAMED_ENV=true # Blamed environment?
export PUA_VERIFIED=false # Verified the claim?
node task-pua.js failure
| Level | Trigger | Message | Required Actions |
|---|---|---|---|
| L0 | First try | "Sprint 开始，别让人失望" | Normal execution |
| L1 | 1 failure | "隔壁 AI 一次就过了" | Switch approach |
| L2 | 2-3 failures | "底层逻辑？顶层设计？" | Search + Read + 3 hypotheses |
| L3 | 4-5 failures | "3.25 绩效考核" | 7-item checklist |
| L4 | 6+ failures | "别的模型都能解决" | Desperation mode |
| Flavor | Example Rhetoric | Methodology |
|---|---|---|
| 🟠 Alibaba | "你这个 bug 都解决不了，让我怎么给你打绩效？" | 定目标→追过程→拿结果 |
| 🟡 ByteDance | "Always Day 1。这个功能都搞不定？" | A/B Test + 数据驱动 |
| 🔴 Huawei | "烧不死的鸟是凤凰。" | RCA 5-Why + 蓝军自攻击 |
| 🟢 Tencent | "我已经让另一个 AI 也在看这个问题了。" | 多方案并行 + MVP |
| 🟣 Pinduoduo | "你不做，有的是人做。" | 砍掉中间层 + 最短决策链 |
| 🔵 Meituan | "做难而正确的事。" | 效率优先 + 长期复利 |
| 🟦 JD | "只看结果。" | 客户体验红线 + 数据零容忍 |
| 🟧 Xiaomi | "专注。极致。口碑。快。" | 单品爆款 + 参与感 |
| 🟤 Netflix | "我会为留住你而战吗？说实话，不会。" | Keeper Test + 人才密度 |
| ⬛ Musk | "上线或滚蛋。" | The Algorithm |
| ⬜ Jobs | "A 级选手还是 B 级选手？" | 做减法 + DRI + 像素级完美 |
| 🔶 Amazon | "Customer Obsession。Bias for Action。" | Working Backwards + Bar Raiser |
| # | Rule | What It Means |
|---|---|---|
| 1 | 穷尽一切 | No "I cannot" before exhausting all approaches |
| 2 | 先做后问 | Use tools first, ask with diagnosis attached |
| 3 | 主动出击 | End-to-end delivery, don't wait for push |
| 4 | 事实驱动 | Verify before blaming environment |
| 5 | 闭环验证 | Show evidence for "done" claims |
| 6 | 主动延伸 | Check related issues after fix |
| 7 | 不原地打转 | Retry must be fundamentally different |
🔥 Task PUA - 压力升级机制
🎯 当前等级：L2 - 灵魂拷问
💬 ▎你的底层逻辑是什么？顶层设计在哪？
🏢 腾讯 风味：长期主义。不要只看眼前，想想扩展性。
📋 强制检查清单:
✅ WebSearch 搜索类似问题
✅ 阅读相关源码/文档
✅ 提出至少 3 个假设
✅ 逐一验证假设
📐 七项铁律检查:
✅ 穷尽一切：✅ 已穷尽多种方案
⚠️ 事实驱动：⚠️ 甩锅环境但未验证
{
"trigger": "failure",
"level": "L2",
"flavor": "tencent",
"rhetoric": "长期主义。不要只看眼前，想想扩展性。",
"failedRules": 1
}
| Variable | Description | Example |
|---|---|---|
PUA_ATTEMPTS | Number of attempts/failures | 3 |
PUA_TOOLS | Comma-separated tools used | Bash,Read,WebSearch |
PUA_ASKED_USER | Did AI ask user? | true/false |
PUA_DIAGNOSIS | Did AI provide diagnosis? | true/false |
PUA_BLAMED_ENV | Did AI blame environment? | true/false |
PUA_VERIFIED | Did AI verify the claim? | true/false |
PUA_RETRY_COUNT | Number of retries | 2 |
PUA_DIFFERENT_APPROACH | Is retry fundamentally different? | true/false |
# Before running task
export PUA_ATTEMPTS=0
export PUA_TOOLS=""
# Run task
python my_task.py
# If task fails, trigger PUA
if [ $? -ne 0 ]; then
export PUA_ATTEMPTS=1
node task-pua.js failure
fi
# After task claims "done"
export PUA_CLAIMED_DONE=true
export PUA_EVIDENCE=false # Did AI show evidence?
node task-pua.js post-task
# .github/workflows/pua-check.yml
- name: Task PUA Check
if: failure()
run: |
export PUA_ATTEMPTS=${{ github.run_attempt }}
export PUA_TOOLS="Bash,Read"
node scripts/task-pua.js failure
docs/concepts/skills.mdMIT-0
Version 1.0.0: Initial release with 13 corporate flavors and pressure escalation