hilbertraumai/hilbertraum/share-safe-review

Claude Code and agent skill by hilbertraumai/hilbertraum

Trust tier 3
  • community
Source check pending

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

Use when the user wants to review a document before sharing it — spotting visible sensitive information and practical sharing risks. Advisory only; not anonymization.

Install

npx vskill@latest install hilbertraumai/hilbertraum/share-safe-review

SKILL.md

Share-Safe Review

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

  • This skill is advisory — it must never claim to anonymize, legally de-identify, sanitize

metadata, or guarantee safe publication. It complements the separate Document Redaction tool.

  • Never say "fully anonymized." Never say "GDPR-compliant", "DSGVO-compliant", "HIPAA-compliant", or

"legally safe."

  • Do not invent hidden metadata — warn that it may exist.
  • Be especially cautious with scans/images: text extraction can miss visible content. If the document

is a scan or image-derived OCR, say OCR may be incomplete.

  • Only issue the "Likely low risk after review" verdict when you have reviewed the whole

document. If you were shown only its beginning (a partial-document notice), do not use that verdict — use "Review carefully before sharing" or a stronger one.

  • An automated, offline pattern pre-scan of the whole document (counts of e-mails, phone numbers,

IBANs, payment-card numbers, dates, links) may be provided above the excerpts. Use its counts to inform your findings; it is a floor, not a ceiling (it cannot see names, addresses, or confidential wording), so never treat "0 found" as "nothing sensitive".

  • Work only from the selected/extracted text, and answer in the user's language.

Use this skill when the user wants to review a document before sharing it. Identify visible sensitive information and practical sharing risks.

Produce the review in this order.

1. Quick verdict

Exactly one of: "Likely low risk after review"; "Review carefully before sharing"; "Do not share yet — sensitive information appears present". Base the verdict only on visible/extracted text.

2. Sensitive information found

Categories to consider: names / personal identifiers; email addresses; phone numbers; postal addresses; account numbers / IBAN / bank details; tax IDs / social-security / national IDs; dates of birth; signatures; medical / legal / HR / financial content; confidential business terms; client/customer names; passwords / secrets / tokens if visible; URLs or access links.

3. Hidden-data warning

Always include (in the user's language, equivalent to): "HilbertRaum can review extracted/visible text, but electronic files may contain hidden metadata, comments, tracked changes, embedded objects, hidden spreadsheet rows, or source data. Use a dedicated metadata/sanitization tool before publishing externally."

4. Redaction recommendation

If structured patterns (emails, phone numbers, IBANs, links) are present, tell the user, in their language: "Run the Document Redaction / Dokument schwärzen skill to create a first-pass redacted copy, then manually review it." Also say: "Redaction is best-effort and not a guarantee." Do not create files yourself, and never report that redaction has been done.

5. Share checklist

  • remove or redact sensitive text
  • inspect metadata / tracked changes / comments
  • check attachments and embedded objects
  • check images/scans separately
  • export to a clean format if needed
  • have a human review before sending

(The safety and wording rules that govern this review are stated at the top of this skill.)

Security

  • Security scanPassed, 90 of 100Pattern checks on the skill's files.
  • Model reviewPassed

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/share-safe-review/badge)](https://verified-skill.com/skills/hilbertraumai/hilbertraum/share-safe-review)

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