All skills
simota avatar

/canon

@e307415
by shingo imotasimota/agent-skills85 stars
15

Assessing standards, regulatory controls, and legal-document coverage with cited evidence and proposed wording. Use for OWASP/WCAG/SOC2/PCI/HIPAA or ToS/privacy/DPA reviews; not legal advice or code fixes.

Use this Skill: https://skilld.dev/gh/simota/agent-skills/canon

This session only. Nothing lands on disk.

referencegdpr-compliance.md

≈2.5k tokens on demand. Your agent reads this file only when SKILL.md points to it.

GDPR Compliance Reference

Purpose: Standards-citation-driven assessment against the EU General Data Protection Regulation (Regulation (EU) 2016/679, effective May 25, 2018). Cites specific Articles, evaluates lawful basis, surfaces DPIA triggers (Art. 35), data-subject rights coverage, breach-notification readiness (Art. 33 — 72 hours), DPO appointment threshold (Art. 37), and routes remediation to specialist agents.

Scope Boundary

  • canon gdpr: standards-citation-driven assessment. Pin "GDPR (Regulation (EU) 2016/679)", cite Article and paragraph (e.g., Art. 6(1)(b), Art. 32(1)(a)), evidence at file:line, lawful-basis mapping per processing activity.
  • cloak (elsewhere): privacy-by-design implementation (Art. 25). PII detection, data-flow maps, consent-management code, DPIA facilitation, anonymization/pseudonymization design. Canon assesses; Cloak builds.
  • canon gdpr / audit / policy: audit trail, Records of Processing Activities (RoPA — Art. 30), Policy-as-Code, DPA registry, and supervisory-authority evidence packages.
  • canon legal-document recipes: legal text review of Privacy Policy, Cookie Notice, and DPA, including clause-gap detection. Canon does not make legal determinations; pair the review with qualified counsel.
  • sentinel (elsewhere): static scanning for PII leakage, hardcoded data-subject identifiers, insecure storage of personal data.

Workflow

SURVEY    →  inventory processing activities, data categories, data subjects, recipients,
             cross-border transfers (Art. 44-49: SCC, BCR, adequacy decision)
          →  pin GDPR + check member-state derogations (e.g., German BDSG, Italian Codice Privacy)

PLAN      →  per processing activity: identify lawful basis (Art. 6 — six bases), special
             category basis (Art. 9 if applicable), purpose, retention, recipients

ASSESS    →  Articles 5/6/7/13/17/25/30/32/33/35 as core; expand to 12-22 (rights), 44-49
             (transfers) as scope demands. Evidence at file:line + DPIA/RoPA refs

VERIFY    →  gap report by Article; flag DPIA-required activities (Art. 35), 72h breach
             readiness (Art. 33), DPO threshold (Art. 37(1)(b)/(c))

PRESENT   →  delegate: Cloak (PbD, DPIA, consent code); continue in Canon (RoPA, DPA evidence),
             Canon legal review (Privacy Policy text), Sentinel (PII scan), Builder (remediation)

Six Lawful Bases (Art. 6(1))

Letter Basis Typical Use Caveat
(a) Consent Marketing emails, optional cookies Must meet Art. 7 — freely given, specific, informed, unambiguous, withdrawable
(b) Contract Account creation, order fulfillment Only data necessary for contract performance
(c) Legal obligation Tax records, AML/KYC Cite the specific law
(d) Vital interests Medical emergency Rare for typical SaaS
(e) Public interest / official authority Public bodies Rare for private sector
(f) Legitimate interests Fraud prevention, internal analytics Requires LIA (Legitimate Interests Assessment); does NOT apply to special categories without separate Art. 9 basis

For special categories (Art. 9 — health, biometric, political, religious, trade union, sex life, racial/ethnic, genetic), requires both Art. 6 basis AND Art. 9 basis (e.g., explicit consent, vital interests, employment law).

Core Articles Assessed

Article Title Assessment Focus
Art. 5 Principles Lawfulness, fairness, transparency, purpose limitation, data minimization, accuracy, storage limitation, integrity/confidentiality, accountability
Art. 6 Lawful basis Per-activity basis declared and documented
Art. 7 Consent conditions Withdrawable as easily as given; pre-ticked boxes invalid
Art. 13/14 Information to data subject Privacy Notice completeness (collected from / collected about)
Art. 15-22 Data-subject rights Access, rectification, erasure (Art. 17), restriction, portability, objection, automated-decision opt-out
Art. 17 Right to erasure ("right to be forgotten") Deletion workflow incl. backups, third-party recipients notification
Art. 25 Data Protection by Design and by Default Default = minimum data, minimum exposure, minimum retention
Art. 28 Processor obligations DPA in place, sub-processor flow-down, audit rights
Art. 30 Records of Processing Activities (RoPA) Maintained for controllers (and processors) — exempt only if <250 employees AND processing not risky/regular/special
Art. 32 Security of processing Encryption, pseudonymization, CIA, resilience, regular testing
Art. 33 Breach notification to supervisory authority Within 72 hours of awareness, unless unlikely to risk rights
Art. 34 Breach communication to data subjects Where high risk to rights and freedoms
Art. 35 DPIA Required when likely high risk; specific triggers below
Art. 37 DPO designation Required for public authority, large-scale special categories, large-scale systematic monitoring
Art. 44-49 International transfers SCC (2021 modules), BCR, adequacy decision (e.g., EU-US DPF), derogations

DPIA Triggers (Art. 35(3) + WP29 Guidelines)

Trigger Example
Systematic, extensive evaluation incl. profiling with legal/significant effects Credit-scoring, insurance underwriting
Large-scale processing of special categories or criminal data Hospital EHR, HR system with health data
Systematic monitoring of public area CCTV, public Wi-Fi tracking
Innovative technology Biometrics, IoT, AI/LLM with personal data
Data combined or matched from multiple sources Marketing data brokers
Data of vulnerable subjects Children, employees (asymmetric power)
Cross-border processing preventing rights exercise Sub-processor in non-adequate country

WP29 rule: 2+ triggers → DPIA required. National DPAs publish their own DPIA-required lists.

Breach Notification (Art. 33) — 72-Hour Clock

  • Clock starts from "awareness" by controller, not from breach occurrence.
  • Notification skipped only if "unlikely to result in a risk to the rights and freedoms of natural persons" — high bar.
  • Phased notification allowed: initial within 72h, details in subsequent updates.
  • To data subjects (Art. 34): only when "high risk" — encryption renders data unintelligible may exempt.
  • Penalties: up to €20M or 4% of global turnover (whichever higher) for Art. 5/6/7/9 violations; €10M or 2% for Art. 33 procedural breaches.

DPO Appointment (Art. 37)

Mandatory when:

  • Public authority/body (except courts in judicial capacity).
  • Core activities = regular and systematic monitoring of data subjects on large scale.
  • Core activities = large-scale processing of special categories or criminal data.

DPO must report to highest management level, cannot be dismissed for performing tasks, and may be staff or contracted external.

Anti-Patterns

  • Treating GDPR as cookie-banner compliance — GDPR governs all personal-data processing; cookies are one slice (and ePrivacy Directive overlaps).
  • Citing "GDPR Article 6" without paragraph — Art. 6(1)(a) consent vs. Art. 6(1)(f) legitimate interests have entirely different obligations. Always cite paragraph + letter.
  • Consent for everything — over-relying on Art. 6(1)(a) when contract or legitimate interests fit better. Withdrawn consent then breaks core service.
  • Paper compliance — RoPA spreadsheet maintained but never updated; Privacy Policy says "we delete on request" but no erasure pipeline exists.
  • Audit-only thinking — annual privacy review without continuous DPIA updates when processing changes (new vendor, new feature, new data category).
  • Copy-paste Privacy Policy from another company — wrong lawful bases, wrong retention, wrong DPO contact. Each Privacy Policy must reflect actual processing.
  • Ignoring sub-processor flow-down — Art. 28(4) requires equivalent obligations on sub-processors; ignoring this leaves controller liable.
  • Treating pseudonymization as anonymization — pseudonymized data is still personal data (Recital 26). Only true anonymization (irreversible, no re-identification) exits GDPR scope.
  • Cross-border transfer on legacy SCCs — 2010 SCCs replaced by 2021 modular SCCs; old contracts after Dec 27, 2022 are invalid.
  • Scope-creep / scope-shrinkage on "controller vs. processor" — misclassification shifts liability and changes obligations (RoPA, breach reporting, DPA).
  • Reactive breach response — discovering on day 70 then panicking. Build a 72h-ready playbook with pre-drafted templates and decision trees.

Handoff

  • To Cloak: privacy-by-design implementation (Art. 25), DPIA facilitation (Art. 35), consent management code (Art. 7), pseudonymization/anonymization design (Art. 4(5)), data-subject rights API (Art. 15-22), data-flow maps.
  • To Canon gdpr / audit / policy: RoPA maintenance (Art. 30), DPA registry (Art. 28), audit trail, OPA rules for retention/access, and supervisory-authority evidence packages.
  • To Canon privacy / cookie / dpa: Privacy Policy / Cookie Notice / DPA legal text review and clause-gap analysis. Always pair with qualified counsel for legal determinations.
  • To Sentinel: PII leakage scanning, hardcoded personal-data identifiers, insecure storage detection, secret/credential exposure (Art. 32(1)(a)).
  • To Crypt: encryption-at-rest/in-transit design, key management for Art. 32(1)(a) and breach-notification exemption qualification.
  • To Builder: technical remediation with Article-citation acceptance criteria (e.g., "Implement Art. 17 erasure pipeline including backups within 30 days").

Source: SKILL.md on GitHub

1 warning13d5 checks · Risk SAFE
  • Gen Agent Trust Hub13d

    The 'canon' skill is a comprehensive framework for assessing software against security, accessibility, quality, and regulatory standards. It provides extensive reference documentation and templates for auditing projects. The skill is verified as safe, with all identified behaviors—such as the use of well-known auditing tools and policy-as-code execution—being standard practices within its functional domain.

  • Socket13d

    No alerts

  • Snyk13d

    Risk: LOW · No issues

  • Runlayer6mo

    2/6 files flagged

  • ZeroLeaks5mo

    Score: 93/100 · 2 sections analyzed

Signed by skilld at e307415. This ties the file your Agent reads to that commit on GitHub. It does not review the instructions.

Last checked against GitHub 2 days ago.

Activeupdated 2 weeks ago

README badge

README badge for simota/agent-skills/canon