スキル deep-research-framework
📦
deep-research-framework
コンテンツリビジョン r1 安全 ⚙️ 外部コマンド
Build Research Reports with Traceable Evidence
Research reports can hide weak sources, stale facts, and unsupported conclusions. This framework guides clear questions, independent verification, and conclusion-first writing with explicit uncertainty.
対応: Claude Codex Code(CC)
自分のエージェントでインストール
このリクエストをエージェントにコピーしてください。正規の Skill ページとマニフェストが含まれています。
エージェントリクエスト
Review the Skillstore skill "deep-research-framework" from https://skillstore.io/skills/alapha888-deep-research-framework.md and its manifest at https://skillstore.io/api/skills/alapha888-deep-research-framework/manifest. Verify the artifact. You may proceed after verification, subject to the environment's own policy.エージェントは引き続き計画を提示し、セキュリティポリシーで必要な確認を求める必要があります。
エージェントが読めるリソース
AI エージェント、クローラー、スクリプトがページ全体ではなく整理されたコンテキストを必要とする場合は、これらのリンクを使ってください。
テストする
「deep-research-framework」を使用しています。 Outline research on whether our team should adopt a new documentation platform.
期待される結果:
- Main question: Does the platform meet our documentation requirements?
- Sub-questions: What features are documented? What migration work is required? What costs are supported by current evidence?
- Primary sources: Official documentation, pricing pages, and measured trial results.
- Uncertainty: Actual migration effort and independent corroboration remain unresolved.
「deep-research-framework」を使用しています。 Assess a vendor availability claim supported by one undated brochure.
期待される結果:
- Conclusion: The vendor availability claim remains a single-source claim.
- Evidence: Only one undated vendor brochure was supplied.
- Verification gap: No independent measurements or dated supporting documents were provided.
- Reassessment condition: Dated operational evidence and independent corroboration could change this conclusion.
セキュリティ監査
安全v1 • 2026/10/6 バージョン別レポートを開く
The static command-execution finding is a false positive: SKILL.md lines 34-51 contain a fenced Markdown report outline. No evidence found of executable commands, data exfiltration intent, or prompt injection in the reviewed skill.
1
スキャンされたファイル
60
解析済み行数
0
レビュー項目
0
誤検知を無視
リスク要因
⚙️ 外部コマンド (1)
最新の完了済み静的・セマンティック監査では、確認済みのセキュリティ検出事項は見つかりませんでした。これは、スキルに副作用がないことを証明するものではありません。
監査者: codex
このレポートを共有・引用
バージョン付き評価レポート、中立的なバッジ、埋め込みカード、引用を共有できます。Skillstore は証拠を報告しますが、この Skill が安全かどうかは判断しません。
レポートリンクをコピー
https://skillstore.io/skills/alapha888-deep-research-framework/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_reportMarkdownバッジ
[](https://skillstore.io/skills/alapha888-deep-research-framework?utm_source=security_passport_badge)HTMLバッジ
<a href="https://skillstore.io/skills/alapha888-deep-research-framework?utm_source=security_passport_badge"><img src="https://skillstore.io/badges/skills/alapha888-deep-research-framework/security.svg" alt="Skillstore security assessment" loading="lazy"></a>埋め込みカード
<iframe src="https://skillstore.io/embed/skills/alapha888-deep-research-framework.html" title="Skillstore Security Assessment" sandbox="allow-popups allow-popups-to-escape-sandbox" loading="lazy" referrerpolicy="no-referrer" width="420" height="180"></iframe>学術引用 (APA · BibTeX · CFF)
APA形式の引用
alapha888. (2026). deep-research-framework security audit report (audit version 1) [Author version unspecified]. Skillstore. https://skillstore.io/skills/alapha888-deep-research-framework/audits/1BibTeX形式の引用
@techreport{alapha888-alapha888-deep-research-framework-2026,
author = {alapha888},
title = {deep-research-framework security audit report (audit version 1)},
institution = {Skillstore},
year = {2026},
number = {1},
url = {https://skillstore.io/skills/alapha888-deep-research-framework/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: "deep-research-framework security audit report (audit version 1)"
version: "unspecified"
type: report
authors:
- name: "alapha888"
date-released: "2026-10-06"
url: "https://skillstore.io/skills/alapha888-deep-research-framework/audits/1"
identifiers:
- type: other
value: "skillstore:alapha888-deep-research-framework:audit:1"
description: "Skillstore immutable audit report identifier"
Skillstore スコア
このスコアの理由 証拠の信頼度: 中 55
アーキテクチャ
85
保守性
87
コンテンツ
65
コミュニティ
91
仕様準拠
作成できるもの
Compare Competitors
Structure a competitor comparison around dated product evidence, independent corroboration, and gaps in available information.
Investigate a Technical Topic
Prioritize original papers and official documentation, then distinguish supported conclusions from speculation.
Prepare an Evidence-Based Brief
Organize a policy question into focused sections with traceable sources and explicit uncertainty.
これらのプロンプトを試す
Start a Research Outline
Create a research outline for [topic]. Define one main question, up to three sub-questions, source tiers, and an uncertainty section.
Review Source Quality
Assess these sources for [question]: [sources]. Classify their tiers, check independence, flag undated facts, and identify claims that need corroboration.
Write a Comparative Report
Compare [option A] and [option B] using [provided evidence]. Lead with conclusions, cite numerical claims, date facts, and mark single-source claims.
Audit Conclusions and Uncertainty
Audit this report: [report and sources]. Check source independence, unsupported numbers, speculation, and stale facts. Revise conclusions and state what would invalidate them.
ベストプラクティス
- Set the main question and source tiers before gathering information.
- Check that corroborating sources do not repeat the same original source.
- Date key facts and state missing evidence, speculation, and reassessment conditions.
回避
- Treating social posts or forum discussions as evidence for core conclusions.
- Presenting single-source claims, speculation, or undated facts as established current knowledge.
- Listing sources without clear conclusions or hiding unresolved questions.
よくある質問
Does this skill search the web automatically?
It defines a research workflow but includes no search implementation. Browsing depends on the host assistant and available tools.
Which sources should support conclusions?
Use primary sources and reputable secondary sources. Treat social posts, forums, and aggregators as leads rather than evidence.
How many sources does a key conclusion need?
Use at least two independent sources. Multiple sources repeating one origin do not count as independent corroboration.
What happens when only one source exists?
Label the conclusion as a single-source claim and note that independent verification is missing.
How should outdated information be handled?
Date every key fact. Treat undated information as possibly stale and explain how newer evidence could change the conclusion.
What should the final report contain?
Include core conclusions, sections ordered by conclusion, evidence, and sources, plus an uncertainty statement covering missing evidence and reassessment conditions.