WorkTwins Product Privacy Supplement

1. Purpose

This supplement explains how WorkTwins specifically handles WorkFootPrints, WorkAtoms, WorkEpisodes, WorkTraces, matching signals, and collaboration features.

2. Core Principle

Raw context is private. Similarity is useful. Users control sharing.

WorkTwins uses work-context signals to calculate computational affinity between users. The purpose is to help compatible people discover each other and collaborate.

3. Similarity-First Processing

WorkTwins is designed to use work signals for comparison, not exposure.

The system does not need to show whether two users match because both held the same object, solved the same error, used the same tool, painted similar objects, or repeated the same workflow. What matters is the compatibility signal.

Whenever possible, WorkTwins transforms raw context into derived WorkFootPrint features, compatibility vectors, workflow categories, embeddings, hashes, confidence scores, and limited match explanations.

4. Local-First, User-Reviewed, Cloud-Assisted

The Subjective Client processes work-context signals locally whenever possible.

Before selected WorkFootPrint information is sent to WorkTwins servers for cloud-assisted matching, the user may review it and choose to:

• approve for matching;
• reduce granularity;
• increase granularity;
• mark private;
• exclude;
• delete;
• request reprocessing;
• view details.

5. Granularity Control

A high-granularity signal may describe a specific workflow, tool, object, task, or problem type.

A lower-granularity signal may describe only a general work category, such as:

• design task;
• debugging workflow;
• writing activity;
• physical construction task;
• collaboration pattern;
• software development task.

Users can reduce granularity for privacy or increase granularity for more accurate matching.

6. Evidence Integrity

WorkTwins is not a CV.

A CV is what people claim they can do. WorkTwins is based on real work signals.

Users may control what is included, excluded, redacted, or shared. However, users cannot rewrite the semantic meaning of detected work evidence to claim that they performed different work.

A user may remove a detected workflow from matching, lower its detail, or mark it private. A user may not change a detected “UI design task” into “backend database architecture” just to improve a profile.

7. Work-Related Signals Only

WorkTwins records work experience, not private life.

WorkTwins does not intentionally record, store, upload, or use non-work-related private activity for matching. This includes adult content, entertainment, personal browsing, private messages, passwords, medical information, financial information, religious activity, political activity, or other unrelated personal activity.

If the Subjective Client detects that activity is not work-related, the system is designed to exclude it from the WorkFootPrint and prevent it from being used for matching.

8. Personal Information Filtering

The Subjective Client and WorkTwins prompts are designed to describe what the user is working on, not unnecessary personal details.

WorkTwins aims to remove or reduce unnecessary personal information and keep only the derived work-context features needed for compatibility comparison.

9. What Is Shared by Default

Raw work evidence is not shared with other users or companies by default.

By default, matching may use derived features such as:

• compatibility vectors;
• workflow categories;
• confidence scores;
• match scores;
• general “matched because” explanations;
• selected profile information.

10. What Other Users May See

Other users may see:

• compatibility score;
• general overlap category;
• limited “matched because” explanation;
• public profile information chosen by the user;
• collaboration availability;
• user-approved evidence cards.

Other users do not receive raw screenshots, private files, source code, passwords, private messages, personal documents, or full local context by default.

11. Company Use

Companies may pay for matching services. Company access does not automatically include raw work evidence.

Company-visible information is limited to configured matching outputs, compatibility summaries, workspace settings, agreement terms, and user-approved sharing choices.

12. Decentralized Direction

WorkTwins currently uses cloud-assisted systems where technically necessary. The long-term direction is decentralized, privacy-preserving matching, including local, federated, or peer-to-peer comparison where technically possible.