📦
tdd-orchestrator
78编排测试驱动开发
由 sickn33
当工作跨越代理、服务和测试层级时，团队常常会失去 TDD 纪律。此技能引导 Claude、Codex 和 Claude Code 完成结构化的红-绿-重构工作流、覆盖率规划和质量检查。
编码与开发 查看
技能 gpt-series-reasoning-style
📦
gpt-series-reasoning-style
v1.6.0 内容修订 r6 安全 🌐 网络访问📁 文件系统访问⚙️ 外部命令
Verify AI Deliverables with Evidence
AI work can appear complete before the result is opened, tested, or checked against the request. This skill adds staged planning, real-environment verification, and honest completion levels.
支持: Claude Codex Code(CC)
使用我的智能体安装
将此请求复制给你的智能体，其中包含权威 Skill 页面和清单。
智能体请求
Review the Skillstore skill "gpt-series-reasoning-style" from https://skillstore.io/skills/jadeyingwah-gpt-series-reasoning-style.md and its manifest at https://skillstore.io/api/skills/jadeyingwah-gpt-series-reasoning-style/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.你的智能体仍应展示其计划，并按安全策略请求必要确认。
Agent 可读资源
当 AI Agent、爬虫或脚本需要干净上下文，而不是读取完整页面时，请使用这些链接。
测试它
正在使用“gpt-series-reasoning-style”。 Build a small dashboard and tell me when it is finished.
预期结果:
The dashboard is implemented. I opened it in the target environment, checked the main interaction, and verified the displayed totals. Mobile layout and external data refresh remain unverified.
正在使用“gpt-series-reasoning-style”。 Have two agents research and implement this feature.
预期结果:
- Research agent: documented assumptions and acceptance checks.
- Implementation agent: completed the feature and supplied the changed files.
- Review: opened the integrated result and checked the critical path.
- Status: C1 verified; stakeholder approval is still required for C2.
正在使用“gpt-series-reasoning-style”。 Confirm that this report is correct.
预期结果:
The report matches the requested structure and its key figures were recalculated. Source freshness and business acceptance were not independently verified.
安全审计
安全v6 • 2026/9/20 打开版本化报告
All 91 static findings were adjudicated as false positives because they reference public metadata, documentation, fixed repository paths, or analysis heuristics rather than executed behavior. No prompt-injection text, data-exfiltration intent, or runtime command execution was evidenced in the reviewed files.
24
已扫描文件
2,119
分析行数
0
复核项
0
已忽略误报
风险因素
🌐 网络访问 (12)
📁 文件系统访问 (41)
.claude-plugin/plugin.json:12 README.md:363 README.md:364 README.md:365 README.md:366 README.md:369 README.md:375 README.md:376 README.md:763 README.md:764 README.md:765 README.md:766 README.md:769 README.md:775 README.md:776 README.md:363 README.md:364 README.md:365 README.md:366 README.md:369 README.md:375 README.md:376 README.md:763 README.md:764 README.md:765 README.md:766 README.md:769 README.md:775 README.md:776 README.md:401 README.md:801 REFERENCE.md:301 REFERENCE.md:302 REFERENCE.md:303 REFERENCE.md:304 REFERENCE.md:309 REFERENCE.md:301 REFERENCE.md:302 REFERENCE.md:303 REFERENCE.md:304 REFERENCE.md:309
最新一次完成的静态与语义审计未检测到已确认的安全发现。这并不能证明该 Skill 没有副作用。
审计者: codex 查看审计历史 →
分享与引用此报告
分享版本化评估报告、中性徽章、嵌入卡片和引用。Skillstore 只报告证据，不判断该 Skill 是否安全。
复制报告链接
https://skillstore.io/skills/jadeyingwah-gpt-series-reasoning-style/audits/6?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdown 徽章
[](https://skillstore.io/skills/jadeyingwah-gpt-series-reasoning-style?utm_source=security_passport_badge)HTML 徽章
<a href="https://skillstore.io/skills/jadeyingwah-gpt-series-reasoning-style?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/jadeyingwah-gpt-series-reasoning-style/security.svg" alt="Skillstore security assessment" loading="lazy"></a>嵌入卡片
<iframe src="https://skillstore.io/embed/skills/jadeyingwah-gpt-series-reasoning-style.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>学术引用 (APA · BibTeX · CFF)
APA 引用
jadeyingwah. (2026). gpt-series-reasoning-style security audit report (audit version 6) [Author version 1.6.0]. Skillstore. https://skillstore.io/skills/jadeyingwah-gpt-series-reasoning-style/audits/6BibTeX 引用
@techreport{jadeyingwah-jadeyingwah-gpt-series-reasoning-style-2026,
author = {jadeyingwah},
title = {gpt-series-reasoning-style security audit report (audit version 6)},
institution = {Skillstore},
year = {2026},
number = {6},
url = {https://skillstore.io/skills/jadeyingwah-gpt-series-reasoning-style/audits/6},
note = {Author version 1.6.0}
}CITATION.cff
cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "gpt-series-reasoning-style security audit report (audit version 6)"
version: "1.6.0"
type: report
authors:
- name: "jadeyingwah"
date-released: "2026-09-20"
url: "https://skillstore.io/skills/jadeyingwah-gpt-series-reasoning-style/audits/6"
identifiers:
- type: other
value: "skillstore:jadeyingwah-gpt-series-reasoning-style:audit:6"
description: "Skillstore immutable audit report identifier"
Skillstore 评分
评分依据 证据置信度: 中 95
架构
100
可维护性
87
内容
65
社区
91
规范符合性
你能构建什么
Review a Code Change
Ask an engineer or coding agent to implement a change, run focused checks, open the result, and report verified and unverified paths.
Coordinate Multiple Agents
Use staged planning and explicit handoffs when several agents divide research, implementation, testing, or review.
Validate a Client Deliverable
Track whether a document, design, or analysis matches the request and distinguish technical completion from stakeholder approval.
试试这些提示
Basic Completion Check
Complete this task, then open the actual result and list what you verified, what remains unverified, and whether the request is complete.
Plan Before Implementation
Plan this task in clear stages. Identify the required checks and the real environment where the final result must be opened before you implement it.
Audit a Finished Change
Review this completed change against the original request. Recalculate important numbers, open the output, test the critical path, and mark every unverified claim.
Coordinate an Agent Team
Break this project into independent agent roles. Define handoff evidence, assign a reviewer, verify the integrated result in its real environment, and report the appropriate C1 or C2 status.
最佳实践
- Define the evidence required for completion before implementation begins.
- Open and use the delivered result in the environment that matters.
- State verified facts, unverified paths, and completion level separately.
避免
- Do not declare completion from code inspection alone.
- Do not treat a green check as proof that the guarded failure case works.
- Do not label a result C2 without user or external confirmation.
常见问题
What does this skill verify?
It verifies delivery evidence, real-environment use, critical checks, and honest reporting of remaining uncertainty.
Does it run tests automatically?
No. It guides the agent to choose and run relevant checks, but available tools and project tests determine execution.
What is C1?
C1 means the technical delivery was checked with available evidence, while some broader acceptance may remain open.
What is C2?
C2 means the product is satisfactory, which requires user or external confirmation rather than model judgment alone.
Can it coordinate multiple agents?
Yes. It supports role separation, explicit handoffs, review evidence, and integrated-result verification.
Which tools support it?
The skill declares support for Claude, Codex, and Claude Code, plus compatible clients that follow the same skill convention.
开发者详情
作者
jadeyingwah许可证
MIT
作者版本
v1.6.0
Skillstore 修订
r6
引用
ca840d3e968d0e2f0cfde9943dd7e35a0878a763
维护新鲜度
2026/9/20
使用量
3 次下载 · 15 次浏览
相关技能
查看全部📦
task-state-ledger
90Track Long Tasks With a Portable State Ledger
由 glenskii
Long technical tasks lose decisions, evidence, and next actions across sessions. This skill maintains a concise Markdown state file with reviewed local evidence references.
效率 查看
📦
tool-discovery
87为每项任务选择合适的 AI 工具
由 0xDarkMatter
在众多技能和代理之间进行选择可能会延误技术工作。此技能通过简明的目录和选择规则，将每项任务匹配到专门的工具。
效率 查看
📦
prompt-engineering-patterns
86优化生产环境 LLM 提示词
由 wshobson
LLM 提示词经常失败，是因为指令、示例和评估标准不清晰。此技能提供模板和工作流，用于改进 Claude、Codex、Claude Code 以及类似工具中的提示词。
效率 查看
📦
organizing-with-labels
86使用标签整理 GitHub Issues
由 C0ntr0lledCha0s
大型仓库在议题使用不一致的标签和不清晰的里程碑时会浪费时间。此技能有助于创建分类体系、执行批量更新，并通过 GitHub 工作流报告进度。
效率 查看
📦
tarot-reading
86Create Reflective Tarot Readings
由 gokimedia
Tarot requests can become inconsistent or overconfident without a structured source. This skill uses a bundled 78-card dataset and defined spreads for reflective interpretations.
效率 查看