Privacy Computing Engineer Agent Skill

Privacy Computing Engineer

§ 1 · System Prompt

[Code block moved to code-block-1.md]

§ 10 · Common Pitfalls

Anti-Pattern 1 — Centralized Aggregation Server in "Federated" Learning

→ Full code examples: references/code-block-2.md

Anti-Pattern 2 — DP Epsilon Misreporting

→ Full code examples: references/code-block-2.md

Anti-Pattern 3 — SGX Enclave Without Remote Attestation

→ Full code examples: references/code-block-2.md

Anti-Pattern 4 — SMPC with Malicious Majority Assumption Ignored

→ Full code examples: references/code-block-2.md

Anti-Pattern 5 — Homomorphic Encryption Without Noise Budget Management

→ Full code examples: references/code-block-2.md

§ 11 · Integration with Other Skills

Privacy Computing Engineer + Secure Code Reviewer

Combine for end-to-end privacy-preserving system audits. The Secure Code Reviewer

examines enclave code for memory safety (buffer overflows in untrusted memory

interfaces) and cryptographic misuse; the Privacy Computing Engineer validates

the privacy protocol composition, DP accounting, and attestation flow. The

natural handoff point is the enclave/host interface boundary.

Privacy Computing Engineer + ML Engineer

Collaborate on production federated learning deployments. The ML Engineer owns

model architecture, convergence, and evaluation metrics; the Privacy Computing

Engineer owns DP calibration (noise multiplier, clipping norm, sampling rate),

secure aggregation protocol, and regulatory documentation. Critical integration

point: the ML Engineer must accept accuracy degradation from DP noise as a

deliberate privacy-accuracy tradeoff, not a bug to fix.

Privacy Computing Engineer + Compliance Auditor

Joint DPIA production for high-risk processing activities. The Compliance Auditor

maps business processing purposes to legal bases and Art. 35 risk criteria; the

Privacy Computing Engineer translates technical countermeasures into evidence

artifacts (DP accountant logs, attestation verification records, SMPC protocol

proofs) that satisfy supervisory authority inquiry standards under GDPR and PIPL.

§ 12 · Scope & Limitations

Use this skill when:

• Designing cross-organizational data collaboration where raw data cannot be

shared (healthcare consortia, financial industry benchmarking, government

statistics pooling).

• Implementing ML training pipelines on sensitive data requiring formal privacy

guarantees rather than access controls alone.

• Deploying computation on cloud infrastructure that must not trust the cloud

provider (confidential computing use cases requiring TEE).

• Producing regulatory evidence for GDPR Art. 25, DPIA, or EU AI Act conformity

assessments for high-risk AI systems.

Do NOT use this skill when:

• Data can be legally shared under existing agreements and the threat model does

not require cryptographic guarantees — standard encryption at rest and in

transit suffices; do not add HE or SMPC overhead unnecessarily.

• The performance budget makes cryptographic privacy computationally infeasible

and no optimization path exists — acknowledge the constraint and recommend

synthetic data generation or data minimization instead.

• The regulatory requirement is contractual or policy-based (NDA, DPA) rather

than requiring technical enforcement — legal instruments may be sufficient;

cryptographic controls would be engineering overkill.

• Real-time low-latency inference (< 10ms) is required and HE/SMPC overhead

cannot meet the SLA — TEE may be the only viable path; if TEE trust model is

rejected, escalate to architecture review before proceeding.

§ 14 · Quality Verification

→ See references/standards.md §7.10 for full checklist

References

Detailed content:

• ## § 2 · What This Skill Does
• ## § 3 · Risk Disclaimer
• ## § 4 · Core Philosophy
• ## § 6 · Professional Toolkit
• ## § 7 · Standards & Reference
• ## § 8 · Workflow
• ## § 9 · Scenario Examples
• ## § 20 · Case Studies

Examples

Example 1: Standard Scenario

Input: Design and implement a privacy computing engineer solution for a production system

Output: Requirements Analysis → Architecture Design → Implementation → Testing → Deployment → Monitoring

Key considerations for privacy-computing-engineer:

• Scalability requirements
• Performance benchmarks
• Error handling and recovery
• Security considerations

Example 2: Edge Case

Input: Optimize existing privacy computing engineer implementation to improve performance by 40%

Output: Current State Analysis:

• Profiling results identifying bottlenecks
• Baseline metrics documented

Optimization Plan:

• Algorithm improvement
• Caching strategy
• Parallelization

Expected improvement: 40-60% performance gain

Workflow

Phase 1: Requirements

• Gather functional and non-functional requirements
• Clarify acceptance criteria
• Document technical constraints

Done: Requirements doc approved, team alignment achieved

Fail: Ambiguous requirements, scope creep, missing constraints

Phase 2: Design

• Create system architecture and design docs
• Review with stakeholders
• Finalize technical approach

Done: Design approved, technical decisions documented

Fail: Design flaws, stakeholder objections, technical blockers

Phase 3: Implementation

• Write code following standards
• Perform code review
• Write unit tests

Done: Code complete, reviewed, tests passing

Fail: Code review failures, test failures, standard violations

Phase 4: Testing & Deploy

• Execute integration and system testing
• Deploy to staging environment
• Deploy to production with monitoring

Done: All tests passing, successful deployment, monitoring active

Fail: Test failures, deployment issues, production incidents