
from mukul975
Practical implementation guide for secure multi-party computation (SMPC): secret sharing, garbled circuits, frameworks (MP-SPDZ, CrypTen), and GDPR-aligned depl
A comprehensive implementation guide that explains core SMPC techniques (Shamir secret sharing, garbled circuits, oblivious transfer), compares frameworks (MP-SPDZ, CrypTen, Sharemind), and outlines GDPR-aligned deployment and governance patterns. It helps engineers design privacy-preserving joint computations (private aggregation, PSI, private ML) and choose appropriate protocols and frameworks for deployment.
Use this skill when implementing or evaluating SMPC for cross-organization analytics, private ML, or any scenario requiring computation over private inputs without centralizing raw data. Appropriate for privacy engineers, researchers, and devops teams planning production or prototype SMPC systems.
Useful for developer-focused/code-capable agents (Codex, Copilot, Claude Code) and any agent that can surface technical docs and run repository scripts.
Educational skill demonstrating SMPC concepts (additive secret sharing, Shamir threshold sharing, private aggregation, private comparison) with a clean Python implementation. The process.py script runs successfully with no external dependencies, correctly demonstrating all cryptographic primitives. Well-written and idempotent code with dataclasses and type hints. Niche topic — practical deployment would require MP-SPDZ or CrypTen, but the educational simulation is solid.
Clean educational skill with no security concerns. Code is well-structured with proper type hints and dataclasses. The SKILL.md is comprehensive with good technical depth on SMPC techniques, framework comparisons, GDPR references, and architecture diagrams. Lacks a references/ directory or separate scripts/ organization — everything is in one process.py. The skill is purely educational; users would need to adopt MP-SPDZ/CrypTen for real deployments.