GPDR-Compliance（GPDR合规工具）
v1.0.3欧盟通用数据保护条例（GDPR）合规专用工具。当用户需要处理欧盟境内的数据保护、 GDPR合规检查、数据保护影响评估（DPIA）、数据主体权利保障、跨境数据传输等相关任务时使用此skill。 **Skill包含的核心工具**： - ✅ `gdpr-check.py` - 基础合规检查脚本 - ✅ `data-su...
⭐ 0· 103·0 current·0 all-time
byWei Wu@wwumit
MIT-0
Download zip
LicenseMIT-0 · Free to use, modify, and redistribute. No attribution required.
Security Scan
OpenClaw
Benign
high confidencePurpose & Capability
Name/description (GDPR compliance, DPIA, data-subject rights, cross-border transfer) align with the included scripts (gdpr-check.py, dpia-generator.py, data-subject-rights.py, cross-border-transfer.py) and templates. Declared dependencies (pandas, jinja2) are reasonable for analysis and template generation. Minor inconsistency: the top-level skill name contains a typo ('GPDR') while package name and code use 'gdpr', but this is cosmetic.
Instruction Scope
SKILL.md promises local-only behavior, only reading references and writing JSON reports; the provided scripts show only local file I/O and template generation (no network libraries in inspected files). One practical caveat: several scripts accept output paths and will create/write files (e.g., save_template, TemplateEngine.save_document) without prompting for explicit user confirmation — SKILL.md states 'all operations need explicit user authorization', but the code does not implement interactive permission checks in non-interactive runs. Also the template engine will create assets/templates directories under the skill path (template_dir.mkdir), which is a benign filesystem write but should be noted as a write operation.
Install Mechanism
No install spec is provided (instruction-only from platform perspective) and the package contains code that runs locally. This is low risk compared to remote-download installers. requirements.txt lists only pandas and jinja2 (well-known PyPI packages). No evidence of downloads from arbitrary URLs or archive extraction in the provided files.
Credentials
The skill requires no environment variables, no credentials, and no special config paths. The dependency list and requested functionality are proportionate to generating reports and running checks locally. No broad credential access or unrelated secret requests were observed.
Persistence & Privilege
Skill metadata flags are standard (always:false, user-invocable:true, model invocation enabled). The skill does not request permanent presence or modify other skills. File writes are local report/template generation; no evidence of modifying system-wide agent settings or other skills' configs.
Assessment
This package appears coherent and local-only, but take standard precautions before use: (1) run it in an isolated/test environment (python venv) and inspect or run the included security_check_gdpr.py as recommended; (2) review any remaining/omitted files (only parts were shown) for unexpected network calls or subprocess usage; (3) when running with real data, back up data and avoid running as an elevated user; (4) be aware scripts will write files to paths you provide and the template engine will create an assets/templates folder — validate output paths to avoid accidental overwrites; (5) remember the tool is informational and not a substitute for legal advice — consult a lawyer or DPO for regulatory decisions.Like a lobster shell, security has layers — review code before you run it.
latestvk975yvs2k8b3qfgf7tep602hzd83s9j3
License
MIT-0
Free to use, modify, and redistribute. No attribution required.