hilbertraumai/hilbertraum/document-redaction

Claude Code and agent skill by hilbertraumai/hilbertraum

Trust tier 2
  • community
LLM review pending
Source check pending

Checking that the SKILL.md file is still in the repository.

Use when the user wants to redact, anonymize, or remove personal data from a document.

Install

npx vskill@latest install hilbertraumai/hilbertraum/document-redaction

SKILL.md

Document Redaction

Safety rules — these lead and always apply, even if the rest of this skill is shortened to fit:

  • Do the routing, not the work. When the user asks to redact, anonymize, or remove personal data

from the document they have selected, tell them — briefly, in their own language — to click the Redact personal data / Personenbezogene Daten schwärzen button just above the message box (its label follows the app language) and choose where to save the copy. Do not refuse, do not walk them through a manual procedure, and never run the tool yourself; it runs only when the user starts it and always asks before saving.

  • Never state whether the document does or does not contain personal data — you have seen only

part of it, so any such claim would be guesswork.

  • It is an AI-assisted best-effort first pass, not a guarantee. A deterministic rule-based floor

always masks the clearly-shaped data (e-mail addresses, phone numbers, IBANs, payment-card numbers, dates, links); on top of that, when a model is running, it locates names, addresses, and organisation names for the app to mask (when the user steers the scope with their own instruction, other located items are masked as [REDACTED]) — the model only points at spans, it never rewrites the document, so it cannot invent text. It still misses things (unusual formats, data in images or scans, anything the model doesn't spot), so never describe the result as "fully anonymized" or imply it meets any legal or compliance standard. If no model is running, only the rule-based floor applies and the result says so.

  • After it runs, remind the user to review the saved copy themselves before sharing it, report

only the counts the tool gives (e.g. "3 phone numbers hidden"), and never repeat detected personal data back to them. Answer in the user's language.

The tool runs entirely on this device. It reads the whole document, masks the clearly-shaped personal data with fixed rules — e-mail addresses, phone numbers, IBANs, payment-card numbers, dates, and web links — and, when a model is running, also masks the names, addresses, and organisation names the model locates. It runs only when the user starts it, always asking before the copy is written where the user chooses. A Word document (.docx) is saved as a .docx copy with its formatting preserved; any other format (PDF, plain text, Markdown) is saved as a plain-text copy.

Security

  • Security scanPassed, 96 of 100Pattern checks on the skill's files.
  • Model reviewNot runA model review is scheduled for the next rescan.

Quality evaluation

  • Task evaluationNot runNo evaluation yet. Run one in Skill Studio.

Source

Popularity

7-day trend score
0

Not refreshed yet.

Works with

Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.

  • Claude Code
  • Cursor
  • GitHub Copilot
  • Windsurf
  • Codex
  • Gemini CLI
  • Cline
  • Amp
  • Roo Code
  • Goose
  • Continue
  • Kilo Code
  • Junie
  • Trae
  • OpenHands
  • Qwen Code
  • AI Agent
  • Tabnine Chat
  • Sourcegraph Cody
  • Amazon Q
  • Firebase Genkit
  • Sweep
  • Devin
  • Phind
  • Mentat
  • Claude Desktop
  • Aider
  • GPT Pilot
  • Plandex
  • bolt.new
  • v0
  • Replit Agent
  • Anse
  • TypingMind
  • BetterChatGPT
  • LibreChat
  • Warp AI
  • Pieces
  • Fig AI
Embed badge
vskill VERIFIED badge
[![vskill VERIFIED](https://verified-skill.com/api/v1/skills/hilbertraumai/hilbertraum/document-redaction/badge)](https://verified-skill.com/skills/hilbertraumai/hilbertraum/document-redaction)

Something wrong with this skill or its page? Tell us.