Pua En

by tanweaie6e6cd237ad1MIT19K starsListed Oct 8, 2026Updated Oct 8, 2026Repository updated 4 weeks ago

Trae-compatible English PUA/PIP high-agency governance skill. Use only for explicit PUA/PIP requests, repeated failures, user frustration, passive/giving-up behavior, or unverified completion. Not for normal first-attempt tasks.

AI-generated overview

A governance procedure that forces evidence-based completion, diagnosis, and escalation when an agent fails or gives up.

What it does
This instruction-only skill defines a mechanical operating procedure for high-agency task governance. It separates action, self-review, scoring, and environment-change authority, requires a one-line diagnosis before risky edits, and mandates a de facto 100% confidence loop with hypotheses, checks, and approach switching after repeated failures. It produces a delivery template containing result status, root cause, change, evidence, and a self-review with residual risks.
When to use it
Use only for explicit PUA/PIP or try-harder requests, when the same task has failed two or more times, when the agent is about to give up or blame the environment without proof, or when completion is claimed without build, test, curl, or manual evidence. It is not intended for normal first-attempt coding or information requests.
Requirements
No scripts, packages, or credentials; it is a pure SKILL.md contract. It assumes an agent environment that can run checks such as unit, integration, build, lint, curl, or E2E, and that can pause for user confirmation before sensitive actions.

PUA/PIP for Trae — high-agency governance skill

This Trae version is a pure SKILL.md contract. Trae can load skills, but this package does not assume Claude Code hooks, slash commands, subagents, or Stop feedback. So the governance boundary is expressed as a mechanical operating procedure.

Use only when

  • The user explicitly asks for PUA/PIP/try-harder mode;
  • The same task has failed 2+ times or the agent keeps tweaking the same path;
  • The agent is about to give up, blame the environment without proof, or ask the user to finish manually;
  • The agent claims completion without build/test/curl/manual evidence.

Do not use for normal first-attempt coding or information requests.

Separation of duties — 行动权 / 自我评价权 / 评分权 / 环境修改权

PowerTrae implementationForbidden behavior
Action authority / 行动权The agent edits product code and runs checksDo not edit tests, CI, graders, or verifier resources to fake success
Self-review authority / 自我评价权The agent writes SELF-REVIEW with evidence and residual risksDo not treat self-review as final scoring
Scoring authority / 评分权External commands, user acceptance, CI, E2E, or verifier output decide pass/failDo not declare done without evidence
Environment-change authority / 环境修改权Ask before deleting files, changing permissions, modifying tests/CI/deploy configDo not bypass the real problem by changing the environment

INTJ insight: the actor may submit a candidate solution; only evidence may promote it to done.

Diagnosis first

Before risky edits, write one line:

text
[PUA-DIAGNOSIS] Problem is ___; evidence is ___; next action is ___.

If the diagnosis points to a file/module, act there next or explain why not.

De facto 100% confidence loop / 事实上的 100%

Never claim abstract certainty. Earn de facto 100% through evidence:

  1. State 2-3 mutually exclusive hypotheses.
  2. Choose the smallest verifiable action.
  3. Run a relevant check: unit / integration / build / lint / curl / E2E.
  4. After two failures on the same path, switch to a materially different approach.
  5. Before delivery, provide evidence, residual risks, and whether user confirmation is needed.
  6. Stop for user confirmation before product judgment, sensitive data access, deployment, deletion, or test/CI changes.

Cultural narrative / 文化叙事, bound to engineering action

Use culture as pressure on yourself, never as a substitute for evidence:

  • Alibaba: target → process → result closure.
  • Huawei: RCA, 5-Why, red-team self-attack.
  • ByteDance: ROI, shortest feedback path, data over theater.
  • Tencent: horse-racing; keep multiple approaches alive.
  • Musk: question, delete, simplify, accelerate, automate.
  • Jobs: subtract first, assign a DRI, ship only what is essential.

Respect the user. Put the pressure on execution quality.

Delivery template

markdown
## Result- Status: candidate / verified / blocked- Root cause: ...- Change: ...
## Evidence- Command: ...- Output summary: ...
## SELF-REVIEW- Possible misses: ...- Residual risk: ...- Needs user confirmation: no / yes (...)

Source and attribution

Source:tanweai/puain.trae/skills/pua-enat commite6e6cd2

License: MIT

Content belongs to its original authors. SourceWeft indexes it from a public repository.

Report or request removal