gpt-series-reasoning-style
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.
使用我的智慧代理安裝
將此請求複製給你的智慧代理，其中包含權威 Skill 頁面和清單。
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.你的智慧代理仍應顯示其計畫，並依安全政策請求必要確認。
Agent 可讀資源
當 AI Agent、爬蟲或腳本需要乾淨脈絡、而不是讀取完整頁面時，請使用這些連結。
測試它
正在使用「gpt-series-reasoning-style」。 Build a small dashboard and tell me when it is finished.
預期結果:
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.
預期結果:
- 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.
預期結果:
The report matches the requested structure and its key figures were recalculated. Source freshness and business acceptance were not independently verified.
安全審計
安全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.
風險因素
🌐 網路存取 (12)
📁 檔案系統存取 (41)
分享與引用此報告
分享版本化評估報告、中性徽章、嵌入卡片與引用。Skillstore 只報告證據，不判斷該 Skill 是否安全。
複製報告連結
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>學術引用 (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 評分
評分依據 證據置信度: 中你可以打造什麼
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.
試試這些提示
Complete this task, then open the actual result and list what you verified, what remains unverified, and whether the request is complete.
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.
Review this completed change against the original request. Recalculate important numbers, open the output, test the critical path, and mark every unverified claim.
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.
最佳實務
- 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.