sure-scale/doc-haus/playbook-dpa
Claude Code and agent skill by sure-scale/doc-haus
- community
Checking that the SKILL.md file is still in the repository.
Firm playbook: positions, fallbacks, and approved clause text for GDPR/UK GDPR data processing agreements — Article 28 processor obligations, international transfers, liability allocation. EU/UK jurisdiction scope.
Install
npx vskill@latest install sure-scale/doc-haus/playbook-dpaSKILL.md
This is an executable firm playbook for data processing agreements under the GDPR and UK GDPR, reviewed from the controller (customer) side. The contract-type and jurisdiction tags for every position below come from this header (GDPR/UK GDPR data processing agreement; EU/UK scope).
Each ## section is one clause type, stated on a fixed schema so the redline tool can paste the firm text byte-exact, the deterministic lint can detect deviation, and document-reviewer can band each clause:
- Preferred — the firm's default opening position. Text inside an
``approved fence is the firm-approved replacement text — copy it byte-exact into the redline tool's replacement` arg.
- Fallbacks — acceptable rungs below Preferred, each introduced by the
"when" condition that justifies offering it (the "offer when" trigger). Text inside an ```approved-fallback fence is the firm's fallback text, to be used only when its stated condition is met.
- Unacceptable — the walk-away floor: the position the firm does not concede
below without a human decision.
- Rationale — why the position sits where it does.
These positions are firm-configurable defaults, calibrated to be market-balanced rather than aspirational, and bounded by the Article 28 GDPR minimums (a statutory floor cannot be overridden downward). The matter's .dochaus/profile.md (see the firm-profile skill) overrides any Preferred, fallback, or above-statutory floor here; where it is silent, these defaults apply. When a counterparty term sits at or below the Unacceptable floor, do not concede: escalate to the firm's privacy / data-protection lead (or the contact named in the profile's Escalation Rules). The escalation is addressed to that role, not a named person, so the position outlasts any individual assignment.
Jurisdiction guard: This playbook applies only to processing subject to the GDPR or UK GDPR. It pairs with the EU pack and the UK packs (EW, SCT, NIR). Do not apply it to a matter with no GDPR/UK GDPR nexus, and flag to the user any DPA whose governing law or processing scope falls outside the EU/EEA and the UK. Where a position below reflects market drafting practice rather than a statutory requirement, the Rationale says so — do not present market backstops (such as fixed-hour breach-notice windows) as legal obligations.
Scope, roles, and documented instructions
Preferred: The processor processes personal data only on the controller's documented instructions, with the Article 28 GDPR mechanics for legally compelled processing and for instructions the processor believes are unlawful.
Unacceptable: a right for the processor to process personal data for its own purposes (including product improvement or model training on identifiable data) outside the controller's instructions; instructions defined solely by the processor's privacy policy as amended from time to time; omission of the duty to inform the controller before legally compelled processing.
Rationale: Documented-instructions language tracks Article 28(3)(a) GDPR, the core of the controller-processor relationship; own-purposes processing rights make the vendor a controller in its own right, which changes the parties' regulatory exposure and must be flagged, not redlined silently.
Confidentiality of processing personnel
Preferred: Every person authorized to process the personal data is bound by confidentiality, contractually or by statute.
Unacceptable: confidentiality limited to employees while contractors and agency staff are unaddressed; confidentiality obligations that expire on termination of employment or of the DPA.
Rationale: The commitment tracks Article 28(3)(b) GDPR; the need-to-know limitation is standard market hardening on top of the statutory floor.
Security measures
Preferred: Appropriate technical and organizational measures anchored to a specific security annex, with a commitment not to materially diminish protection during the term.
Unacceptable: security obligations described only by reference to the processor's website policy as amended at the processor's discretion, with no non-diminishment commitment; measures qualified by "commercially reasonable" in place of the risk-appropriateness standard; no security annex or equivalent identifiable baseline at all.
Rationale: The standard tracks Article 32 GDPR; the annex anchor and the non-diminishment covenant are the market mechanics that make it enforceable (the Common Paper DPA likewise anchors security to a referenced security policy rather than leaving it free-floating).
Subprocessors
Preferred: General written authorization tied to a published subprocessor list, advance written notice of changes with an objection right, full flow-down of DPA obligations, and processor liability for subprocessor performance.
Fallbacks: when the processor's notice mechanism is a subscription-based mailing list or feed rather than direct written notice, and the objection and flow-down mechanics are otherwise intact.
Unacceptable: unrestricted subprocessor engagement with no notice or objection mechanism; an objection right whose only outcome is that the controller continues under protest; flow-down stated as "substantially similar" obligations without the no-less-protective standard; disclaimer of processor liability for subprocessor acts and omissions.
Rationale: Notice, objection, flow-down, and retained liability are the Article 28(2) and 28(4) GDPR package; the ten-business-day notice period matches the Common Paper DPA position, and the termination-on-unresolved- objection mechanic is the market remedy that gives the objection right teeth.
Assistance with data subject requests
Preferred: The processor assists the controller with data subject rights by appropriate technical and organizational measures, routes any requests it receives to the controller, and does not respond on its own.
Unacceptable: assistance conditioned on uncapped, undisclosed fees set at the processor's discretion; a right for the processor to respond to data subjects directly without the controller's instruction; assistance limited to access requests while erasure, objection, and portability are unaddressed.
Rationale: Tracks Article 28(3)(e) GDPR. Reasonable cost-recovery for extraordinary assistance is market-acceptable, but only the controller is positioned to decide how a rights request is answered.
Personal data breach notification
Preferred: Notification without undue delay after the processor becomes aware of a personal data breach, with a 72-hour contractual backstop, required content, and ongoing cooperation.
This is the start of the file. Read the full SKILL.md on GitHub.
Security
- Security scanPassed, 96 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingdochaus/playbooks/playbook-dpa/SKILL.md
Popularity
- Repository stars
- 80
- Repository forks
- 22
- 7-day trend score
- 0
Stars and forks are for the whole GitHub repository, not this skill alone. 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
[](https://verified-skill.com/skills/sure-scale/doc-haus/playbook-dpa)