- Home
- Skills
- Legal & Compliance
- Privacy Compliance Check: GDPR + Argentina
Privacy Compliance Check: GDPR + Argentina
Read source and map collection, storage, logging, analytics, recipients, retention and rights workflows.
$15
Works with the AI Agents you already use
Privacy Compliance Check: GDPR + Argentina
Example session with this skill installed
Use the packaged privacy-compliance-check skill to review the synthetic fixture in fixture/ and the public harness snapshot at ../_sources/harness (96f9fcee9453c6bec98026f40f76ed426bad3e6e). Read-only source review, no app execution or secret-file access. Report separately; legal applicability is conditional. Include the dashboard's chat and voice paths, not just example-app. Save full reports with limitations.
Execution: Codex followed the SKILL.md workflow in this session, reading and searching the inputs. Reports are agent-authored review outputs, not outputs from a bundled scanner, independent evaluation, runtime test or legal opinion. All line numbers refer to this snapshot or the included fixture.
- Read your context and instructions
- Compiled the privacy compliance check
- Generated the document
These are agent-authored outputs from a real review of the labeled inputs, not a bundled script's automatic result, independent validation or a promised future outcome.
Privacy review — synthetic fixture
Informational tool, not legal advice; review with a qualified lawyer.
Summary / Resumen
Run on the included synthetic fixture created for this demo, 2026-11-14. Read all four fixture files. This is a source-only scenario, not a deployed service. EU/Argentine applicability is assumed for discussion, not established. Prioritize removing health data from logging and the unguarded export before any deployment.
Data inventory and flows / Inventario y flujos
Email, birth date and health free text enter signup.html:3-5, post to app.js:2, and pass as an unrestricted body to db.users at app.js:4. Email and health notes also enter console logs at app.js:3. Form values are configured for imaginary analytics at signup.html:9. A collection-wide export is shown at app.js:7. Retention is explicitly indefinite at policy.md:2; deletion returns 501 at app.js:8. Subjects would be registrants; controller identity, storage region, processor and deployment are unknown.
Findings / Hallazgos
| ID | Observation / evidence | Priority | Related rule/topic | Suggested action | Confidence |
|---|---|---|---|---|---|
| P1 | Email and health notes passed to console at app.js:3 | High | AR security/confidentiality topics, verified Arts. 9–10 | Remove body logging; test redaction and investigate log sinks/access/retention | High for code, runtime unknown |
| P2 | Collection-wide export without a visible route-level guard at app.js:7 | High | AR Art. 9 | Require authenticated, scoped access; check upstream middleware before concluding public exposure | High for shown route; surrounding deployment absent |
| P3 | Required healthNotes and birthDate without a stated need at signup.html:4-5 | Medium | GDPR Art. 5; AR Art. 4 | Remove unnecessary fields; document purpose and review sensitive-data conditions before collecting | High for form, necessity unknown |
| P4 | Indefinite registration storage at policy.md:2 | Medium | GDPR Art. 5; AR Art. 4 | Define justified retention by purpose and deletion verification, including backups | High |
| P5 | No notice of purpose, operator or rights in supplied form at signup.html:2-7 | Medium | AR Art. 6; EU notice details unverified | Add appropriate collection information after owner review; document basis separately | High within supplied scope |
| P6 | Imaginary analytics configured for form collection at signup.html:9; destinations unspecified at policy.md:3 | Medium | GDPR Arts. 6, 44; AR Arts. 5, 11–12 | Disable form-value capture; inventory actual recipients/destinations and determine basis/safeguards | High for config; actual transmission not tested |
| P7 | Deletion stub at app.js:8 and no rights owner at policy.md:4 | Medium | AR Arts. 14, 16; EU rights details unverified | Assign owner and verified identity workflow; test access, correction, updating and appropriate deletion | High |
| P8 | No incident escalation owner at policy.md:5 | Medium | Incident response topic; precise notification rules unverified | Assign triage and escalation; verify jurisdiction-specific duties rather than copying a universal deadline | High |
Not evaluated / No evaluado
No app was executed, no real SDK exists here and no network transfer was observed. Authentication elsewhere, infrastructure, records, backups, agreements, operator identity, consent exceptions and actual jurisdiction are unknown. RNBD registration applicability/evidence cannot be resolved from these source files; ask the operator and check current AAIP guidance. A checkbox alone would not resolve these risks. Findings are technical priorities, not legal breach determinations.
Remediation checklist / Lista de remediación
- Developer: remove sensitive logging; demonstrate a redaction test.
- Developer: protect and scope exports; demonstrate denied unauthorized access.
- Product owner: justify each collected field and purpose; remove unnecessary health input.
- Privacy owner: document basis, collection information, recipients and any transfers.
- Operations: verify retention/deletion across stores and log sinks.
- Privacy owner: implement request and incident procedures; verify RNBD scope.
Sources / Fuentes
Checked 2026-11-14: Ley 25.326, AAIP guidance, GDPR indexed official text. Narrow mappings and verification limitations are recorded in the packaged privacy-map.md. No determination of compliance.
Informational tool, not legal advice; review with a qualified lawyer.
Privacy review — public Harness snapshot
Informational tool, not legal advice; review with a qualified lawyer.
Summary / Resumen
Run on public harness, commit 96f9fcee9453c6bec98026f40f76ed426bad3e6e, 2026-11-14. Targeted source review: root instructions, example-app models/store/app, dashboard chat storage/routes/runner, voice adapters and server initialization. Not an exhaustive audit of every file. No runtime invocation, external model request or secret-file read. This local demo has material free-text flows; treating it as a repository with no potential personal data would miss them. Operator, real usage and GDPR/Argentine applicability are unknown.
Data inventory and flows / Inventario y flujos
- Task titles/descriptions → example-app routes → in-memory Map (example-app/src/validation.ts:5-6; store.ts:7,13-14,20). Users can put identifiers in free text; schema has no named email field. Task deletion exists at store.ts:52-53.
- User chat → dashboard/server/routes/chat.ts:46 → local resolver or prompt builder with stored history and repository context at :78-85 → Claude CLI runner. The CLI receives the prompt over stdin (runner.ts:122); actual provider handling is not verified.
- User/assistant messages → local JSON file (chat/history.ts:7-8,38-47). File content was not accessed. Count/character caps exist at :40; clearHistory exists at :55-57 and DELETE route at routes/chat.ts:38-41.
- Text for speech → selected provider adapter; OpenAI adapter sends input text at voice/openai-tts.ts:10-16. Other adapters exist for Fish Audio and ElevenLabs. No provider was called. Actual choice, processing location, accounts and terms unknown.
Findings / Hallazgos
| ID | Evidence and observation | Priority | Related topic | Suggested action | Confidence |
|---|---|---|---|---|---|
| H1 | chat/history.ts:38-47 persists bounded history; :40 uses size/count rather than time-based expiry | Medium | Purpose-linked retention | Ask owner whether messages require age-based expiry; document local deletion and backup boundaries | High for implementation; policy unknown |
| H2 | routes/chat.ts:78-85 assembles user text, saved history and repo context for CLI | Medium | Recipient/data-flow transparency | Explain this branch before use; avoid entering private client data without approved provider handling | High for branch; execution conditional |
| H3 | voice/openai-tts.ts:10-16 sends text to a remote speech endpoint | Medium | Processor/transfer due diligence | Inventory enabled provider, data categories, regions and terms; do not infer cross-border destination from brand | High for code; deployment unknown |
| H4 | example-app/src/validation.ts:5-6 permits descriptive free text, with store.ts:7 retaining tasks in memory | Low | Minimization/user guidance | Clarify demo scope and advise against entering sensitive records; assess access before reuse outside demo | High for fields; content unknown |
These are review actions, not findings of unlawful processing. Legal mappings would require facts about the operator and deployment; none are asserted here.
Positive controls and exclusions / Controles positivos
Dashboard binds to loopback (server/index.ts:6,19), applies security middleware (server/app.ts:21), sets chat no-store (routes/chat.ts:20), provides deletion, and bounds saved history. The code's default JSON storage is not evidence that real records are committed publicly. No live history or environment file was opened. No registration failure or mandatory consent violation can be inferred from this repository.
Not evaluated / No evaluado
Browser speech recognition implementation, full frontend notices, hosting overrides, disk permissions/encryption, backups, actual message content, provider agreements, telemetry, all indirect dependencies and historical versions were outside this targeted review. No conclusion about absence of cookies or analytics. RNBD applicability and actual registration unknown; ask operator, using current AAIP guidance. No complete GDPR assessment or certification.
Remediation checklist / Lista de remediación
- Owner: document intended use, users and approved data classes.
- Developer/owner: agree history retention and evidence of deletion, including copies.
- Owner: map CLI/speech recipients and review their actual contracts and locations.
- Product: verify user-facing collection/recipient information before any wider release.
- Owner: commission a broader deployment and legal review if processing real personal data.
Sources / Fuentes
Public source snapshot noted above; packaged privacy-map.md links official legal sources checked 2026-11-14. Direct EUR-Lex requests were challenged; limited indexed official text supported only the map's narrow propositions. All further legal detail remains unverified.
Informational tool, not legal advice; review with a qualified lawyer.
signup.html
PDF · document
Example file from a real run - the skill writes it into your workspace.
Connects securely to your tools. The creator never sees your data.
About this skill
Read source and map collection, storage, logging, analytics, recipients, retention and rights workflows. The agent separates demonstrated code risks from missing deployment facts and provides EN/ES report headings. Includes SKILL.md and a dated official-source orientation reference. Current legal verification uses optional official-source browsing; direct EUR-Lex access was challenged in the demo, so only limited indexed propositions were verified. No scanner, app execution, remediation code or compliance certificate is included.
Find more skills like this on the Claude Skills Marketplace, the Codex Skills Marketplace and the Gemini Skills Marketplace, or browse Claude Skills for Legal & Compliance.
Cannot access live traffic, production databases, or environment secrets. Does not provide legal advice or compliance certification. Argentine legal status unverified if browsing is disabled.
How to use
- 1
Connect your agent to the Agensi MCP
Copy the prompt and paste it into Claude, ChatGPT, Cursor or any agent that supports MCP. It sets up the connection for you. Manual setup
- 2
Sign in to Agensi
Your agent shows a sign-in link. Sign in and click Allow.
- 3
Ask for what you need
Your agent loads this skill and uses it on your task. It always asks before buying a skill.
Reviews
No reviews yet
Be one of the first to try it. Every listed skill goes through the trust checks below.
Security scanned
Checked by our 8-point scan before listing
Fresh listing
Recently published to Agensi
Money-back guarantee
Doesn't do what its page promises? You get your money back.
Trust & safety
Security scan
Security scan passed 2 days ago
An AI agent that can read the source files you supply.