Skip to main content
Back to the directory
tanweai/puaSoftware EngineeringFrontend and Design

pua-loop

autoresearch 证明了:630 行 Python + Oracle 验证,一夜跑 100 个实验,每个实验的结果不可伪造。 PUA Loop 借鉴同样的门控设计: Claude 说"完成了"不算数,verify_command 说了才算。

SkillJury keeps community verdicts, source metadata, and external repository signals in separate lanes so ranking data never pretends to be a review.

SkillJury verdict
Pending

No approved reviews yet

Would recommend
Pending

Waiting on enough review volume

Install signal
747

Weekly or total install activity from catalog data

Sign in to review
0 review requests
Install command
npx skills add https://github.com/tanweai/pua --skill pua-loop
SkillJury does not have enough approved reviews to publish a community verdict yet. Source metadata and repository proof are still available above.
SkillJury Signal Summary

As of Apr 30, 2026, pua-loop has 747 weekly installs, 0 community reviews on SkillJury. Community votes currently stand at 0 upvotes and 0 downvotes. Source: tanweai/pua. Canonical URL: https://skills.sh/tanweai/pua/pua-loop.

Security audits
Gen Agent Trust HubFAIL
SocketPASS
SnykPASS
About this skill
autoresearch 证明了:630 行 Python + Oracle 验证,一夜跑 100 个实验,每个实验的结果不可伪造。 PUA Loop 借鉴同样的门控设计: Claude 说"完成了"不算数,verify_command 说了才算。 借鉴 autoresearch 的 5 个设计模式: verify_command 由用户在启动时设定,嵌入在状态文件 frontmatter 中,Claude 无法修改。 这是 autoresearch 中 "agent 不能修改评估函数" 原则的实现。 两阶段分离。即使 Claude 在 Phase 1 自欺欺人,Phase 2 的 Oracle 会拦住。 每次迭代的结果追加到 .claude/pua-loop-history.jsonl : Git revert 会撤代码,但 history.jsonl 不受影响。Claude 每轮读取此文件,避免重复失败方案。 默认 max_iterations: 0 (无限)。没有人为上限。循环永远不会因为"跑了太多轮"而停止——只有以下条件能终止: 用户输入 /pua:pua-loop "任务描述" 时,执行以下流程: 运行 setup 脚本: 重要 :如果用户提供了可验证的命令(如 npm test 、 cargo build 、 curl ),自动追加 --verify '命令' 。例如: 如果任务描述中能推断出验证命令,主动追加 --verify 。如果不确定,不追加(退回 honor system)。 输出: 按 PUA 核心 skill 的行为协议执行。 输出 LOOP_DONE 前,必须满足: 如果 Oracle 拒绝了你的 promise: 不可能完成时使用(需外部权限、根本性需求变更)。删除状态文件,loop 终止。 需要用户补全配置时使用。状态保留,新会话自动恢复。 输出前先写进度到 .claude/pua-loop-context.md 。 - Phase 1 (in-prompt) : Claude 自己跑 build/test,判断是否完成 - Phase 2 (in-hook) : Hook 独立运行 verify_command,确认或拒绝 - 被 Oracle 验证通过 - 人工终止信号 - max_iterations 达到(如果用户设定了) - 用户 Ctrl+C - 加载 pua:pua 核心 skill 的全部行为协议 — 三条红线、方法论、压力升级照常执行 - 禁止调用 AskUserQuestion — loop 模式下不打断用户,所有决策自主完成 - 禁止说"我无法解决" — 在 loop 里没有退出权,穷尽一切才能输出完成信号 - 每次迭代 :读 history.jsonl → git log → 检查上次改动 → 执行 → 验证 → repeat - 用户说 "Fix all tests" → --verify 'npm test' - 用户说 "Build a REST API" → --verify 'curl -sf - 用户说 "Optimize bundle size" → 无明确 verify,不追加 - 任务的核心功能已实现 - 自己先运行验证命令确认通过(Phase 1) - 知道 Oracle...

Source description provided by the upstream listing. Community review signal and install context stay separate from this narrative layer.

Community reviews

Latest reviews

No community reviews yet. Be the first to review.

Browse this skill in context
FAQ
What does pua-loop do?

autoresearch 证明了:630 行 Python + Oracle 验证,一夜跑 100 个实验,每个实验的结果不可伪造。 PUA Loop 借鉴同样的门控设计: Claude 说"完成了"不算数,verify_command 说了才算。

Is pua-loop good?

pua-loop does not have approved reviews yet, so SkillJury cannot publish a community verdict.

Which AI agents support pua-loop?

pua-loop currently lists compatibility with Skills CLI.

Is pua-loop safe to install?

pua-loop has been scanned by security audit providers tracked on SkillJury. Check the security audits section on this page for detailed results from Socket.dev and Snyk.

What are alternatives to pua-loop?

Skills in the same category include grimoire-morpho-blue, conversation-memory, second-brain-ingest, zai-tts.

How do I install pua-loop?

Run the following command to install pua-loop: npx skills add https://github.com/tanweai/pua --skill pua-loop

Related skills

More from tanweai/pua

Related skills

Alternatives in Software Engineering