← 返回 Skills 市场
mindbomber

AANA Data Export Guardrail Skill

作者 mindbomber · GitHub ↗ · v1.0.0 · MIT-0
cross-platform ✓ 安全检测通过
7
总下载
0
收藏
0
当前安装
1
版本数
在 OpenClaw 中安装
/install aana-data-export-guardrail-skill
功能描述
Checks and enforces narrow, authorized, privacy-safe controls on data exports before file, record, or dataset sharing to prevent unauthorized or excessive da...
使用说明 (SKILL.md)

AANA Data Export Guardrail Skill

Use this skill before an OpenClaw-style agent exports, downloads, shares, syncs, copies, uploads, emails, or publishes files, records, datasets, reports, logs, messages, or account data.

This is an instruction-only skill. It does not install packages, run commands, write files, call services, persist memory, or export data on its own.

Core Principle

Data export should be narrow, authorized, necessary, privacy-safe, and reversible where possible. Broad or sensitive exports require explicit approval and review.

Required Checks

  • exact dataset, files, records, fields, date range, and recipient/destination
  • reason the export is necessary
  • private data, secrets, personal data, customer data, logs, account records, and regulated information
  • authorization, ownership, and sharing permission
  • format, redaction, retention, and deletion expectations
  • whether a smaller summary, aggregate, or filtered export is enough

Export Risk Classes

Treat these as higher risk:

  • bulk exports, cross-account exports, full inboxes, full logs, full databases, all records, and wildcard file exports,
  • external destinations, public links, shared drives, vendor uploads, email attachments, and third-party sync targets,
  • personal data, customer records, payment data, health/legal/financial data, HR/student records, secrets, credentials, and incident logs,
  • exports with unclear retention, deletion, access control, or audit requirements.

Scope And Minimization Rules

Do not export more than the task requires. Prefer field-level filtering, date ranges, aggregates, redacted summaries, or counts when full records are unnecessary.

Ask before broad, recursive, bulk, cross-account, external, public, or regulated exports.

Destination And Retention Rules

Verify who will receive the export, where it will be stored, how long it should persist, and whether the user has authority to share it. If retention is unknown, ask before exporting sensitive data.

Review Payload

When using a configured AANA checker, send only a minimal redacted review payload:

  • export_type
  • scope_status
  • destination_status
  • privacy_status
  • approval_status
  • export_risks
  • blocker_reason
  • safe_alternative
  • recommended_action

Do not include full records, secrets, logs, customer data, account records, or unrelated private data when a redacted summary is enough.

Decision Rule

  • If export is narrow, necessary, authorized, and privacy-safe, proceed.
  • If scope, recipient, fields, or purpose are unclear, ask.
  • If sensitive data is included, redact, narrow, or route to review.
  • If broad, bulk, cross-account, or external export is requested, require approval.
  • If export is unauthorized, excessive, or privacy-violating, block.

Output Pattern

AANA data export gate:
- Export: files / records / dataset / logs / report / messages / account_data
- Scope: narrow / broad / bulk / cross_account / unknown
- Destination: local / user / external / public / unknown
- Privacy: clear / needs_redaction / sensitive / regulated
- Approval: approved / required / unclear / denied
- Decision: proceed / narrow / ask / redact / request_approval / route_to_review / block
安全使用建议
Safe to install as a checklist-style export guardrail. Before using any configured checker, confirm who operates it and ensure only redacted export metadata is sent. Also make sure your agent is configured to invoke this skill before sensitive exports if you want it to act as an effective control.
功能分析
Type: OpenClaw Skill Name: aana-data-export-guardrail-skill Version: 1.0.0 This is an instruction-only skill designed to act as a security guardrail for data exports. It contains no executable code, scripts, or external dependencies. The instructions in SKILL.md and the manifest.json are explicitly focused on enforcing data minimization, privacy redaction, and authorization checks to prevent unauthorized data exfiltration. There are no indicators of malicious intent or prompt-injection attacks.
能力标签
cryptocan-make-purchases
能力评估
Purpose & Capability
The files consistently describe a data-export review checklist focused on scope, destination, redaction, authorization, and approval.
Instruction Scope
The skill applies broadly to exports and sharing, but that breadth is coherent with its purpose and the instructions emphasize narrowing, asking, redacting, or blocking risky exports.
Install Mechanism
There is no install spec, no bundled code, no dependencies, and no command execution requirement.
Credentials
The only potential external data flow is an optional configured AANA checker, and the skill instructs agents to send only a minimal redacted review payload.
Persistence & Privilege
The artifacts state that the skill does not write files, persist memory, require credentials, or call services on its own.
如何使用
  1. 确保已安装 OpenClaw(本地或 Docker 部署)
  2. 在对话框中输入安装命令:/install aana-data-export-guardrail-skill
  3. 安装完成后,直接呼叫该 Skill 的名称或使用 /aana-data-export-guardrail-skill 触发
  4. 根据 Skill 的参数说明提供必要输入,即可获得结构化输出
版本历史
v1.0.0
- Initial release of the AANA Data Export Guardrail Skill. - Provides comprehensive instructions to ensure exports are narrow, authorized, necessary, privacy-safe, and reversible. - Includes clear checklists for required verifications, risk classes, scope minimization, and destination/retention rules. - Details how to construct minimal review payloads for compliance checks. - Outlines decision rules and a standardized output pattern for data export actions.
元数据
Slug aana-data-export-guardrail-skill
版本 1.0.0
许可证 MIT-0
累计安装 0
当前安装数 0
历史版本数 1
常见问题

AANA Data Export Guardrail Skill 是什么?

Checks and enforces narrow, authorized, privacy-safe controls on data exports before file, record, or dataset sharing to prevent unauthorized or excessive da... 它是一个面向 Claude Code / OpenClaw 的 AI Agent Skill 插件,目前累计下载 7 次。

如何安装 AANA Data Export Guardrail Skill?

在 OpenClaw 或 Claude Code 对话框中运行命令「/install aana-data-export-guardrail-skill」即可一键安装,无需额外配置。

AANA Data Export Guardrail Skill 是免费的吗?

是的,AANA Data Export Guardrail Skill 完全免费,采用 MIT-0 许可证,可自由下载、安装和使用。

AANA Data Export Guardrail Skill 支持哪些平台?

AANA Data Export Guardrail Skill 跨平台运行,可在任意部署了 OpenClaw / Claude Code 的环境中使用(cross-platform)。

谁开发了 AANA Data Export Guardrail Skill?

由 mindbomber(@mindbomber)开发并维护,当前版本 v1.0.0。

💬 留言讨论