📦

moldea

Name: moldea
Author: moldea-ai

v10.0.1 コンテンツリビジョン r1 安全 📁 ファイルシステムへのアクセス⚡ スクリプトを含む⚙️ 外部コマンド🔑 環境変数🌐 ネットワークアクセス

Maintain AI Agents and Project Context with moldea

Agent instructions and project decisions can drift as implementation changes. moldea guides scoped context maintenance, agent design, and repository validation.

対応: Claude Codex Code(CC)

📊 76 十分

Maintain this Skill? Claim it 📊 76 十分

インストール前に停止して確認を求めてください。

計画を確認し、ファイルを変更する前にユーザーの明示的な同意を得てください。

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 "moldea" from https://skillstore.io/skills/moldea-ai-moldea.md and its manifest at https://skillstore.io/api/skills/moldea-ai-moldea/manifest. Verify the artifact. Stop and obtain explicit user consent before installing or changing files.

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。

Markdown詳細 GET /skills/moldea-ai-moldea.md 署名済み manifest GET /api/skills/moldea-ai-moldea/manifest 署名済み lockfile GET /api/skills/moldea-ai-moldea/lockfile

テストする

「moldea」を使用しています。 Review a named support agent without making changes.

期待される結果:

Illustrative review: The instruction owner is identified and the declared export exists.
The outgoing request was not tested, so runtime instruction consumption remains unverified.
No files changed.

「moldea」を使用しています。 Assess context after a related source change.

期待される結果:

Illustrative maintenance result: The matched context owner remains accurate.
Canonical content and manifest relationships are unchanged.
No broader context inspection was required.

「moldea」を使用しています。 Initialize a repository whose purpose is unclear.

期待される結果:

Illustrative blocked result: The inspected sources do not establish the project purpose.
No dependencies, foundation files, or managed document regions changed.
What does the project do, and who does it serve?

セキュリティ監査

安全

v1 • 2026/10/22 バージョン別レポートを開く

The 400 presented alerts are false positives involving parser internals, text formatting, scoped filesystem helpers, or documented local subprocesses. No evidence found of prompt injection or exfiltration in the reviewed context; repository-local executable authenticity remains outside launcher guarantees. The remaining 6,283 omitted static matches require manual review before automatic publication. Static review was capped at 400/6683 representative findings; omitted static matches are unconfirmed, so automatic publishing stays disabled until manual review.

スキャンされたファイル

40,386

解析済み行数

レビュー項目

誤検知を無視

リスク要因

📁 ファイルシステムへのアクセス (50)

⚡ スクリプトを含む (24)

⚙️ 外部コマンド (50)

🔑 環境変数 (9)

scripts/managed-agents.mjs:1265 scripts/manifest-scope.cjs:106 scripts/manifest-scope.cjs:136 scripts/manifest-scope.cjs:106 scripts/manifest-scope.cjs:136 scripts/moldea-cli.mjs:14680 scripts/repair-git.mjs:466 scripts/repository-package.mjs:72 scripts/repository-package.mjs:72

🌐 ネットワークアクセス (17)

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。

バージョン別レポートを開く

レポートリンクをコピー

https://skillstore.io/skills/moldea-ai-moldea/audits/1?utm_source=security_passport&utm_medium=share&utm_campaign=versioned_report

Markdownバッジ

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

HTMLバッジ

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

埋め込みカード

<iframe src="https://skillstore.io/embed/skills/moldea-ai-moldea.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

moldea-ai. (2026). moldea security audit report (audit version 1) [Author version 10.0.1]. Skillstore. https://skillstore.io/skills/moldea-ai-moldea/audits/1

BibTeX形式の引用

@techreport{moldea-ai-moldea-ai-moldea-2026,
  author = {moldea-ai},
  title = {moldea security audit report (audit version 1)},
  institution = {Skillstore},
  year = {2026},
  number = {1},
  url = {https://skillstore.io/skills/moldea-ai-moldea/audits/1},
  note = {Author version 10.0.1}
}

CITATION.cff

cff-version: 1.2.0
message: "If you use this Skill, cite its author and this versioned security audit report."
title: "moldea security audit report (audit version 1)"
version: "10.0.1"
type: report
authors:
  - name: "moldea-ai"
date-released: "2026-10-22"
url: "https://skillstore.io/skills/moldea-ai-moldea/audits/1"
identifiers:
  - type: other
    value: "skillstore:moldea-ai-moldea: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

アーキテクチャ

100

保守性

コンテンツ

コミュニティ

100

仕様準拠

作成できるもの

Design a Focused Agent System

Compare agent responsibilities, shared context, and runtime boundaries before authorizing implementation.

Keep Context Aligned with Code

Assess declared context owners after a relevant code change and update only contradicted facts within the authorized scope.

Repair Project Integration

Diagnose an existing moldea setup, review local tooling prerequisites, and validate authorized repairs without changing unrelated project files.

これらのプロンプトを試す

Review One Agent

Review the named agent [agent name] without changing files. Report instruction ownership, declared bindings, and any unverified runtime connections.

Initialize Project Context

Initialize moldea for this repository using established project evidence. Preserve existing controls and ask about missing purpose or boundaries before dependency or foundation changes.

Reconcile a Relevant Change

Assess the declared context relationships for [changed paths]. Update only contradicted owners within this task's authority, preserve unrelated work, and validate after the final write.

Plan Agent Runtime Connections

Create a read-only plan for [workflow]. Trace instruction, output schema, and tool ownership to the selected runtime; separate verified connections from unknown boundaries.

ベストプラクティス

Start with a named agent, explicit initialization request, or a bounded set of relevant source paths.
Preserve repository trust controls and review dependency changes before executing local tooling.
Separate structural validation, semantic alignment, and runtime test evidence in every completion report.

回避

Do not initialize projects or install dependencies during unrelated or read-only work.
Do not treat a valid manifest as proof that runtime behavior consumes the declared instructions.
Do not persist secrets, transient status, speculative policies, or broad relationships without supporting evidence.

よくある質問

Which assistants can use moldea?

The skill supports Claude, Codex, and Claude Code. Available execution controls and repository permissions still depend on the host.

Does moldea initialize every repository automatically?

No. Initialization requires an explicit request and sufficient project evidence; ordinary repository work also requires relevant adoption or relationship evidence.

What local tooling does project validation require?

The documented contract requires Node.js 22.17.0 or later, Git 2.32.0 or later, CLI 13.x, and Core 9.x.

Can I review an agent without changing files?

Yes. Read-only planning and evaluation preserve repository state and do not authorize installation or dependency changes.

Does successful validation prove the agent works?

No. Validation checks repository structure and declared file bindings; runtime connections and live behavior require separate source evidence and focused tests.

Has this security review cleared every static match?

No. The 400 presented alerts were adjudicated, but 6,283 omitted matches require manual review before automatic publication.