agentleFS
nassaj2011nassaj2011/claude-skills-for-chatgpt/skills/privacy-data-governance/SKILL.md

privacy-data-governance

Skill0 starsChanged 26 days agoMITView on GitHub

Trace personal-data flows, minimization, user rights, retention, vendor transfers, deletion propagation and privacy release evidence for software platforms.

The file

We split this skill into the rules it gives an agent

2 rules. About 0 are common, 2 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈2

    Distinctive · estimate

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    If repository, storage, analytics configurations or tools are unavailable, ask for redacted schema/data-flow descriptions and produce an …

    line 28
  2. Distinctive:

    No covert telemetry, hidden source-code collection, secret extraction or backdoor.

    line 34

≈ “Common” is an estimate: at least 1 in 5 of the 364,980 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Privacy and data governance — executable review contract

Use when

The task involves personal data, logs/analytics, identity, marketing preferences, uploads, vendors, data exports or customer rights. Do not invoke for a purely static CSS change without user data.

Evidence-first workflow

  1. Scope and authority. Identify jurisdictions, controller/processor relationships, products, audiences, data subjects, environments and actual sources available. Label unknowns; do not assert GDPR or other legal compliance from generic guidance. Obtain qualified legal review for legal basis, cross-border transfers, sensitive classes and statutory timelines.
  2. Data map by field, not by feature name. Record each field and purpose, source, processing system, recipient, access role, transmission/storage region, legal-basis question, retention trigger, deletion path and evidence reference. Include analytics, monitoring, test snapshots, exports, support tickets, message vendors and backups; distinguish operational logs from business records.
  3. Classify each disclosure. Identify required vs optional fields, security/PII severity and whether user choice is required. Apply data minimization and pseudonymization at both collection and each recipient boundary; avoid raw phones, IDs or free-text payloads in analytics/traces.
  4. Map lifecycle. Describe access/correction/export/deletion/withdrawal requests, system-of-record, processors, downstream copies, caches, archives, backups and restoration risk. Never promise instant erasure from immutable backups; state actual retention/restore quarantine controls and lawful exceptions as decisions.
  5. Tie risks to controls. Specify owner, severity, exact remediation, source of verification, test case and blocker status. Prioritize unauthorized disclosures or unclear external sharing before cosmetic policy text.
  6. Verify on permitted environments. Negative tests must cover role- and tenant-scope reads, analytics payload schema, CSV export permissions, deletion propagation, processor contracts, restored-backup deletion replay and incident logging without raw PII. Use synthetic records only.
  7. Decide and hand off. Mark each gate PASS / FAIL / NOT RUN / NEEDS LEGAL REVIEW. Do not call legal compliance 'verified' solely because a checklist is complete.

Required deliverable

  • Data-flow inventory: field | source | purpose | recipient | basis to verify | region | retention trigger | deletion path | evidence.
  • Risk register: severity | affected data | exploit/exposure pathway | owner | corrective action | acceptance test | status.
  • Lifecycle and user-rights map: entrypoint, identity verification, response/exceptions, cascade, backup behavior.
  • Concrete acceptance tests with an expected result; note tests not run and legal questions unresolved.
  • Executive decision to proceed, block, or defer with explicit accountable sign-off, not a legal conclusion.

Browser-only / ChatGPT web

If repository, storage, analytics configurations or tools are unavailable, ask for redacted schema/data-flow descriptions and produce an example matrix with NOT VERIFIED labels. Provide SQL or tests to run later rather than implying access. Never upload, store, echo or transmit real customer datasets unnecessarily.

Verified local reference

An illustrative field-inventory validator and redacted sample live in examples/privacy-inventory/ in the public source checkout. They implement project-defined data policy checks, not legal certification. Run its tests only if a permitted local Python environment exists; otherwise use the checklist above.

Safety constraints

No covert telemetry, hidden source-code collection, secret extraction or backdoor. Do not invent legal deadlines, consent grants, data-controller roles, contractual rights or privacy compliance. Never export real personal data into a public skill repository.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.