Sensitive File Anonymization
Paste the following prompt into your AI chat to install this skill:
Please install @user_c0e4b78f/anonymize-sensitive-files according to the guide at https://skillhub.cn/install/skillhub.md.
About this skill
Problem
When sharing documents with external teams, teams often need to preserve structure while removing names, phone numbers, account identifiers, and keys. Plain search-and-replace can miss hidden fields in DOCX, Excel, and PDF files, or replace the same entity inconsistently across locations.
How It Works
The Skill runs locally through a scan, anonymize, and verify workflow. By default it creates redacted copies instead of overwriting source files:
- Run scan first to inspect warnings, skipped_inputs, and finding_counts before handling high-risk files.
- Use --manifest during anonymization so the same original value maps to the same pseudonym in a single run.
- Run verify with the same rules, vocabulary, and manifest to confirm there are no residual findings or unprocessed inputs.
- Format-specific handling covers DOCX, Excel, and PDF; PDF uses real redaction rather than visual black boxes.
Boundaries
It does not guarantee coverage of every artifact: comments, charts, macros, scanned pages, image text, complex formulas, non-UTF-8 text, and external links may require manual review. If a field rule misses a value, add a failing test first, update the rule, then rerun scanning, anonymization, and verification.
Use Cases
- Before delivering customer contracts, redact names, account IDs, and keys in DOCX, Excel, and PDF while preserving structure.
- Before publishing research results, sanitize sensitive fields in CSV, JSON, and YAML files and generate verifiable redacted copies.
- When investigating false positives or misses, inspect scan warnings and finding_counts, then rerun anonymization and verify with updated field rules.
- When handing off project materials, use the manifest to keep the same original value mapped consistently and export a controlled mapping for later checks.
Best For
- Business or legal assistants who deliver contracts, quotes, and research reports need to replace sensitive fields with stable pseudonyms.
- Analysts handling customer datasets need redacted copies of CSV, JSON, and YAML files without overwriting originals.
- Automation or agent engineers need to connect scanning, anonymization, and verification into Python 3.10+ workflows.
- Data security engineers auditing outputs need to check warnings, residual findings, and unprocessed inputs before release.
Related Skills
An AI workflow that structures personal knowledge bases, IMA, Yuque, and Feishu content into approved personas for paid WeChat mini-agent deployment.
Local workflow memory with matching and SOP updates.
An OpenClaw live streaming executor that initializes TRTC streaming, starts a real-time dashboard, generates viewer URLs, and continuously reports live events.
Breaks down physical supply chains for super-trends to identify second- and third-layer bottlenecks, runs valuation and reverse checks, and maintains trackable reports.