技能 executor-spec

📦

executor-spec

Name: executor-spec
Author: atri10

內容修訂 r1 安全 ⚙️ 外部命令🌐 網路存取📁 檔案系統存取

Write Testable Specifications Before Planning

Ambiguous requirements create conflicting implementations and unverifiable reviews. This skill guides linked specifications, risk assessments, and verification strategies through explicit human approval.

支援: Claude Codex Code(CC)

📊 75 尚可

Maintain this Skill? Claim it 📊 75 尚可

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。

智能體請求

Review the Skillstore skill "executor-spec" from https://skillstore.io/skills/atri10-executor-spec.md and its manifest at https://skillstore.io/api/skills/atri10-executor-spec/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。

Markdown 詳情 GET /skills/atri10-executor-spec.md 已簽名 manifest GET /api/skills/atri10-executor-spec/manifest 已簽名 lockfile GET /api/skills/atri10-executor-spec/lockfile

測試它

正在使用「executor-spec」。 Draft acceptance requirements for placement requests that name an unknown cell.

預期結果:

Requirement R02: A request naming an unavailable cell returns HTTP 422 with the error value unknown_cell.
Observable: The response contains the required status and error value; no placement row is added.
Verification: One criterion checks the rejection response, and another checks the absence of a stored placement.
Review state: Draft awaiting explicit human approval.

正在使用「executor-spec」。 Assess placement latency risks before approving the specification.

預期結果:

Failure scenario: Placement latency exceeded the agreed threshold as tenant count increased.
Cause: The available-cell scan was not measured at the expected tenant scale.
Mitigation: Add a measurable latency requirement and a linked load-test criterion.
Evidence plan: Record the observed percentile latency at the specified tenant count; keep the criterion pending until execution.

正在使用「executor-spec」。 Change a requirement threshold after implementation plans already depend on the specification.

預期結果:

Change classification: Material because the new threshold can change implementation and review outcomes.
Document action: Create a replacement specification and preserve the previous specification body.
Traceability: Keep surviving requirement numbers and relink the verification strategy to the replacement.
Plan impact: Record which dependent plans need revision and which remain unaffected.

安全審計

安全

v1 • 2026/10/5 開啟版本化報告

Most findings are Markdown formatting, legitimate test examples, or fixed sibling references, not shell substitution, reconnaissance, or exploitable traversal. The verification example writes to predictable /tmp/b, creating a local symlink overwrite risk if copied and executed. No evidence found of credential exfiltration or prompt injection; referenced sibling helpers are outside the reviewed package.

已掃描檔案

1,175

分析行數

複核項目

已忽略的誤報

能力複核項目 (1)

這些是真實的本機能力，對此技能而言可能是預期行為，因此需要複核，但不會按已確認的惡意行為計分。

中

references/verification-template.md:47

Temp directory access

| V02 | R02 unknown cell rejected | integration | `curl -s -o /tmp/b -w '%{http_code}' localhost:808

curl writes to predictable /tmp/b without exclusive creation. Another local user could pre-create a symlink and redirect the write to an accessible file.

風險因素

⚙️ 外部命令 (50)

🌐 網路存取 (1)

references/verification-template.md:69

📁 檔案系統存取 (16)

references/verification-template.md:47 SKILL.md:116 SKILL.md:129 SKILL.md:130 SKILL.md:131 SKILL.md:213 SKILL.md:231 SKILL.md:352 SKILL.md:360 SKILL.md:514 SKILL.md:603 SKILL.md:666 SKILL.md:667 SKILL.md:668 SKILL.md:669 SKILL.md:514

審計單位: codex

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。

開啟版本化報告

複製報告連結

https://skillstore.io/skills/atri10-executor-spec/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdown 徽章

[![Skillstore security assessment](https://skillstore.io/badges/skills/atri10-executor-spec/security.svg)](https://skillstore.io/skills/atri10-executor-spec?utm_source=security_passport_badge)

HTML 徽章

<a href="https://skillstore.io/skills/atri10-executor-spec?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/atri10-executor-spec/security.svg" alt="Skillstore security assessment" loading="lazy"></a>

嵌入卡片

<iframe src="https://skillstore.io/embed/skills/atri10-executor-spec.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

atri10. (2026). executor-spec security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/atri10-executor-spec/audits/1

BibTeX 引用

@techreport{atri10-atri10-executor-spec-2026,
  author = {atri10},
  title = {executor-spec security audit report (audit version 1)},
  institution = {Skillstore},
  year = {2026},
  number = {1},
  url = {https://skillstore.io/skills/atri10-executor-spec/audits/1},
  note = {Author version unspecified}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "executor-spec security audit report (audit version 1)"
version: "unspecified"
type: report
authors:
  - name: "atri10"
date-released: "2026-10-05"
url: "https://skillstore.io/skills/atri10-executor-spec/audits/1"
identifiers:
  - type: other
    value: "skillstore:atri10-executor-spec:audit:1"
    description: "Skillstore immutable audit report identifier"

Skillstore 評分

評分依據證據置信度: 中

架構

可維護性

內容

社群

規範符合度

你可以打造什麼

Prepare an implementation specification

Convert approved design candidates into numbered requirements, constraints, scope exclusions, and linked acceptance criteria.

Review delivery risks before approval

Connect failure scenarios to requirements, mitigations, accepted risks, and unresolved questions before authorizing planning.

Define auditable acceptance evidence

Map requirements and constraints to named tests, manual steps, negative cases, boundary checks, and observable outputs.

試試這些提示

Draft a first specification

Use executor-spec to draft requirements from [approved design handoff] for [initiative]. Include constraints, exclusions, and owned questions. Stop for human review before planning.

Add a linked risk assessment

Review [draft specification] with executor-spec. Write a failure narrative, trace causes, assign mitigation links, and document accepted risks. Update affected requirements before approval.

Complete verification coverage

Create a verification strategy for [specification]. Cover every requirement and constraint with exact procedures and expected observations. Include rejection and boundary checks. Keep results pending.

Supersede a specification with dependent plans

Assess [proposed change] against [specification] and [dependent plans]. For material changes, preserve existing IDs, draft supersession documents, relink verification, and record plan impacts.

最佳實務

Provide the approved design handoff, charter, architecture records, and interface documents before drafting requirements.
Use exact thresholds and observable outcomes, then map every requirement and constraint to verification evidence.
Review companion helpers before execution, use private temporary files, and require explicit approval before passing the phase gate.

避免

Replace missing decisions with vague terms such as appropriate, robust, or performant.
Claim tests passed when the document only defines expected evidence.
Edit material requirements in place after dependent plans exist, or treat silence as approval.

常見問題

Which AI tools can use this skill?

The report lists Claude, Codex, and Claude Code as supported tools.

What documents does this workflow produce?

It guides a specification, a risk assessment, and a verification strategy with linked document and requirement identifiers.

Does the package include the Executor helper scripts?

No. The reviewed package contains the skill instructions and three templates; companion scripts and contract references must be provided separately.

Does creating verification criteria mean the tests passed?

No. Criteria remain pending until verification executes their procedures and records observed evidence.

Can an approved specification be changed?

Material changes require supersession once dependent plans exist. Surviving requirement numbers remain stable, and affected plans receive impact records.

Are there precautions before running example commands?

Review companion helpers and adapt commands to your repository. Replace the predictable temporary file with a private temporary file before execution.