Skip to content
Skills Library
Security · Planning

data-classification

/data-classification

Label data by sensitivity so controls, retention, and access follow the label instead of being argued case by case. Use when designing storage, granting access, or deciding how carefully a dataset must be handled.

Repository
AI-SKILLS ↗
Licence
MIT
Last repository change
05.09.2026

Assessment and security

90/ 100Class: good
Format25 / 25
Frontmatter with name and description, valid name, description up to 1,024 characters, name matching the folder.
Discoverability25 / 25
Description explains when to use the skill, has a reasonable length and uses third person.
Content25 / 35
At least 150 words and at most 500 lines, headings, step lists, examples or code.
Hygiene15 / 15
No placeholders such as TODO, no Windows paths, with a licence.
Riskno signals detected in the scan
Files in this package
1
Scripts
0
HTML comments
0
URLs in SKILL.md
0
Deprecation

No signals detected. Repository changed on 05.09.2026, without references to retired models or APIs.

Licence

MIT, open: use and modify while retaining author attribution.

What it does and when to use it

Label data by sensitivity so controls, retention, and access follow the label instead of being argued case by case. Use when designing storage, granting access, or deciding how carefully a dataset must be handled.

Original author description

Label data by sensitivity so controls, retention, and access follow the label instead of being argued case by case. Use when designing storage, granting access, or deciding how carefully a dataset must be handled.

How it works

  1. Method
  2. Boundaries
Original SKILL.md headings · 41 lines, 285 words
SKILL.md preview
---
name: data-classification
description: Label data by sensitivity so controls, retention, and access follow the label instead of being argued case by case. Use when designing storage, granting access, or deciding how carefully a dataset must be handled.
---

# Data classification

Without labels, every dataset gets the same treatment, which means
either over-protecting cheap data or under-protecting sensitive data.
Classification exists so the control follows the content
automatically rather than depending on who is looking.

## Method

1. **Use few levels with clear tests.** Public, internal, confidential,
   and restricted is enough for most organisations, and each level
   needs a one-line test anyone can apply without a meeting.
2. **Classify at the field level where it matters.** A table is rarely
   uniformly sensitive; an email column and a status column do not
   deserve the same handling. Field-level labels let you protect
   precisely (see data-minimization).
3. **Attach controls to the label, not the system.** Encryption,
   access approval, logging, retention, and export rules should be
   defined once per level and applied wherever that level appears (see
   data-encryption, audit-logging).
4. **Label at creation and carry it downstream.** A classified field
   copied into a warehouse, an export, or a log keeps its
   classification, and pipelines must propagate the label rather than
   dropping it at the first join (see data-lineage).
5. **Treat special categories separately.** Health, biometrics,
   financial account data, government identifiers, and children's data
   carry rules of their own beyond your internal scheme.
6. **Review labels when purpose changes.** Data aggregated or joined
   can become more sensitive than either input, so re-classify at those
   boundaries rather than inheriting the lowest label.

## Boundaries

- Classification directs controls; it does not implement them. A label
  with no enforcement is documentation, not protection.
- Over-classification is a real cost: if everything is restricted,
  people route around the controls to get work done.
- Regulatory categories do not always map onto internal levels, so keep
  the mapping explicit rather than assumed.

Versions

RepositoryScoreLicence
This page · Amey-Thakur/AI-SKILLS90MIT
seyhunak/skilldeck100MIT
kaandgnsh/Claude-Blue100MIT

Database snapshot: 13.11.2026 · package version c0fe498e7681 (SHA-256 digest of the entire package)

Downloads in the last 30 days: 0 · ranking

Report a problem with this skill: harmful content, incorrect descriptions or an author’s removal request.

Similar skills

  • 100data-classification-assistant
    Classifies data assets by sensitivity level (Public, Internal, Confidential, Restricted) and generates handling guidelines, retention policies, and access control recommendations. Validates classification against organizational policy and flags misclassifications. Use when asked…
  • 89data-security-analysis
    Analyze data security events, sensitive information type (SIT) access, sensitivity label access, DLP matches, or Purview insider risk activity. Triggers on keywords like "data security", "sensitive information type", "SIT access", "DLP events", "DataSecurityEvents", "EDM…
  • 90data-classification-to-dlp-protocol
    Use this skill when sensitive data must be discovered, classified with Microsoft Purview sensitivity labels, protected by Data Loss Prevention policies, and monitored for label adoption and DLP policy effectiveness across Microsoft 365 and Power Platform environments. Defines…
  • 90authorization
    Enforce authorization checks for external data source access in AI systems. Use when designing, configuring, or reviewing access controls between AI components and external databases or data services.
  • 100data-classification-for-ai
    Defines how the organisation's classification scheme and sensitivity labels govern AI - which classes may ground AI, be typed into prompts, reach which models or external services, or train and tune models - plus how labels carry into AI outputs and how unlabelled data is…
  • 85iam-access-management
    Grants least-privilege access to systems and cloud resources through roles rather than individual permissions, short-lived credentials, and regular review, so standing access doesn't accumulate unnoticed. Use this whenever the user is granting cloud or system access, designing…

Feedback and reports

Your message goes to Bartosz Sokoliński. An email address is needed only for a reply. See the privacy policy.