# OLD DRAFT

use https://hackmd.io/18l9wPFdSxaa4LkcPQRpKA instead

# Verifiable Presentation of Composed Credentials

## Co-Authors
* Diana
* Fateme
* Golda
* Ivan
* Markus
* Sebastian

# Abstract

This paper introduces a novel framework for the presentation of sets of related verifiable credentials that would apply for different purposes; such as infection protection, donation, or education.The framework allows individuals to assemble multiple credentials issued by different issuers and disclose them all together as one claim to stakeholders. By composing credentials together, we can extend the usual pattern of issuer, holder, and verifier. The further challenge is to generate a credible presentation that will be compelling to stakeholders such as regulators, donors, or employers. By bridging the gap between stakeholders and the world of verifiable credentials, we provide a practical and adaptable solution for a wide range of use cases. This holistic exploration of composite credentials demonstrates their potential to address and change the way credentials are presented and verified in a variety of settings.

# Introduction

The digital transformation of credentials has been a subject of interest for both academia and industry, especially in the context of verifiable credentials. Verifiable credentials are digital statements made by the issuer about a subject, which can be verified by anyone without contacting the issuer. Traditionally, the model of verifiable credentials has been relatively straightforward, involving three main roles: the issuer, the holder, and the verifier. The issuer provides the credential, the holder stores it, and the verifier checks its validity. This model works well for single-purpose credentials but falls short when individuals need to present a set of related credentials for diverse use-cases.

Evaluating the truth or validity of the digital statement is a separate matter and may benefit from context, such as having access to statements about the issuer, other independently signed statements relating to the subject or to the statement itself.

Hence, we propose several methods of presenting composed credentials, that may provide valuable context for the evaluation of the a credential.

# Background information and Related works

## Concepts

#### Verifiable Presentation of multiple credentials

This concept refers to the practice of bundling several credentials together in a single package for the purpose of verification. Instead of verifying each credential individually, the bundled set can be verified all at once, making the process more efficient.

#### Composing two or more linked credentials

##### Linking Credentials by Common Subject

This involves creating a unified set of credentials that relate to a single individual or entity, known as the subject. The aim is to provide a comprehensive view of that subject's qualifications, skills, or statuses.

##### Linking Addressable Credentials

Addressable credentials have unique identifiers that allow them to be accessed individually. The concept of linking addressable credentials involves bringing such credentials together, even if they are stored in different locations or issued by different organizations, for the purpose of verification.

##### Credential as a Subject

In some scenarios, a credential itself can serve as a subject for another credential, creating a hierarchical relationship between them. This adds an additional layer of context to the credentials, making them more informative.

##### Linking other Digital Evidence by Hash

This concept refers to the inclusion of additional forms of digital evidence, such as documents or media files, in the credential verification process. These additional pieces of evidence are linked to the credentials through cryptographic hashes to ensure their integrity.

#### Verifiable Attestations on Credentials

An attestation is an extra layer of verification added to a credential by a third party. This concept enhances the credibility of the credential by providing additional information or endorsements that can be useful during the verification process.

#### Public sets of Verifiable Credentials

This concept involves making a set of credentials publicly available for anyone to verify. This is often used in scenarios where transparency is important, such as public registries or open-source projects.

## Use cases

### Food
The food industry sector is regulated and businesses are obligated to proof conformance with legal standards. One legally required standard is to have all emplyees instructed on hygiene regulations according to the infection protection regulation. The instruction itself is conducted by a municipal institution which certifies the successful participation with the respective "health-instruction"-credential. The employee presents this credential on its first commencing day to the first employer, which confirms the receipt by countersigning the credential. The "health-instruction"-credential issued to the employee in combination with this "countersignature"-credential of the first employer will be verified by all future inspections. Even if the employee switches to a new employer, the initial credential combination is verified. The combined proof may be presented by the employee or by the current employer, thus a technical concept is needed to have the two credentials combined by either role when presenting the proof. Several technical implementations are possible:
1. The credentials are stored in wallets and are not adressable. In this scenario the credentials must be physicaly exchanged several times between the wallets of the employer and the employee by messaging provided by the means of the respective ssi-ecosystem. This handling is cumbersome and involves a lot of manual interaction
2. The credentials are addressable such that either role may act on the same instance of the credential.

In this paper we examine how these requirements may be covered by adressable combined credentials.

### Education

Another use case of this idea could be applied to vocational training and education. In the education sector, students often collect various certificates, diplomas, and awards from different schools or training programs. Traditionally, each of these credentials has to be presented individually to employers or educational institutions, making the process cumbersome and time-consuming. Our new framework aims to simplify this by allowing students to combine all their credentials into a single, easy-to-share package.

In this use case, we have three essential roles: *Issuer*, *Holder* and *Verifier*. As the name suggests, the Issuer is responsible for issuing the VC, usually an authorized organization or institution who provides some courses offers, issues and signs the credential. The Holder receives his VC from the Issuer and stores it in his own storage, gaining full control over it. He can decide to show his credential to the Verifier, who now can verify the credential by validating the signature that the Issuer gave upon creation. In this context, the issuer would be an educational institution, the holder is the student, and the verifier could be an employer or another educational institution.
The framework for the presentation of sets of binded or related VCs in this use case could be as follow:

1. Educational Institution (Issuer): May issue one or more credentials for various skills or courses that the student has passed.
2. Learner (Holder): Aggregates his credentials (VC1) and presents them as a composite credential to their employer.
3. Employer (Verifier1): Investigates the validity of binded VCs as a composite VC. They may issue an additional VC (VC2) that signifies the learner’s new role or skill level within the company.
4. Education/Industry Regulatory Body (Verifier2): Checks the validity of the new issued VC (VC2), binded to the previous VCs to ensure that the validity of those VCs.

In this workflow, the flexibility to assemble multiple credentials (possibly from different issuers) into a single, composite credential allows for a more holistic representation of an individual’s skills, experiences, and competencies.

### Philanthropy
In philanthropy it is particularly important to verify that funds are distributed for the intended purpose and to the intended recipients. Generally there are three parties in philanthropy: the funder, the organization providing benefits, and the beneficiaries. Currently, the organization is likely to provide some detailed reporting to the funder, and possibly a third party auditor may be engaged to verify the information in the report.

Verifiable credentials have the potential to give agency and leverage to the recipients themselves, by allowing them to directly attest about the benefits received. If the recipients are also otherwise credentialled, by peer recommendations, work performance or academic attestations, this can add to the overall credibility of the collected set.

In areas where corruption is rampant, having an independently verifiable set of related attestations about use of funds, may make a critical difference in how they are actually distributed.

# Methodology

### Presenting Multiple Credentials
Existing methods allow the construction of a Verifiable Presentation that contains multiple signed credentials, that may have a common subject. The Holder who signs the Verifiable Presentation is commonly the subject of each credential but is not required to be.

### Addressable Verifiable Credentials
In some cases, it may be desirable to publish Verifiable Credentials to a public or privately shared dataset. For example, a restaurant's safety certification may be public, while the individual workers' food handling credentials are held privately and only presented to inspectors.

Several data protocols could be good candidates for publishing verifiable credentials in an addressable manner. In a related paper, we demonstrate how the Ceramic protocol can be used to publish verifiable credentials in a to a decentralized network, thus making them addressable.

Note that addressable verifiable credentials may not be fully public, as encryption or access controls can also be put in place.

### Verifiable Presentations of Addressable Credentials
In order to fulfill the requirements of a Verifiable Presentation, the presentation must be signed by a Holder.

In the case of the Ceramic Network, the Holder is the owner of the data stream. A proof may be included that the Holder signing the Verifiable Presentation is also the DID that is the Controller of the stream where it was published.

This aspect is novel and may be expanded on in a separate paper.

### Linking Verifiable Credentials in a Presentation
In the case where the credentials do NOT have a common subject, it may be valuable to show a proof of the link between them.

Using Addressable credentials allows one credential to actually be the Subject of another, in that a second person can attest to the validity of a claim.

This can be a powerful way of validating a specific claim without having to layer on business logic.

For example, if a student claims that they took a specific course on specific dates, they may make a self-attestation to that effect. A professor could make a similar attestation, but it would be necessary to compare the course dates and name in order to ascertain that the professor is referring to the same course. However if the professor simply directly "vouches for", or attests that *the student's claim is true* then there is no need to compare field by field. This is only possible if the student's claim is addressable.

One way to do this is to have all of the claims embedded in the same document, which gives them relative addresses, but this requires tight coordination. Having each claim have an independent and verifiable address allows the attestations to happen asychronously and even enables adversarial attestations.

### Privacy concerns with Addressable Claims
Note that it is not necessary for a claim to be on a public network for it to be addressable. Several layers of privacy are possible with addressable claims:

1) Wallet address
Generally Verifiable Credentials are requested from the holder by type. It is possible to construct an address starting with the DID of the holder and followed by the type and/or id of the verifiable credential, and to provide a library that converts such a URI to a DIDComm or other standard request method. This allows addressing the credential but it may be difficult to ensure consistency in constructing the address; it might not be determinate.

2) Encryption
Even on a public network, data may of course be encrypted for privacy, and the keys provided only as desired. The tools for constructing the Verifiable Presentation may include selective decryption of fields the Controller or Holder wishes to share.

3) Access Control
It is also possible to have a Ceramic or other network on private infrastructure, rather than on the shared mainnet.

# Possible workflows and processes

In the realm of verifiable credentials, the workflows and processes can vary significantly depending on the use case, the stakeholders involved, and the types of credentials being used. Below, we outline some possible workflows and processes that could be employed to make the most out of the framework for presenting composed credentials.

## Single Holder, Single Verifier Workflow

In the simplest scenario, a single holder presents a set of composed credentials to a single verifier. The verifier checks the validity of each credential in the set and makes a decision based on the aggregated information. This workflow is straightforward but limited in scope, as it doesn't take advantage of the full potential of composed credentials.

## Multi-Issuer, Single Holder, Single Verifier Workflow

In this workflow, a user may have several credentials from multiple issuers and combines them into a single presentation. The verifier then checks the validity of the composed set, which could provide a more comprehensive view of the holder's qualifications or status. This is particularly useful in scenarios like job applications or academic admissions where multiple aspects of an individual need to be assessed.

### Delegation Workflow

In a delegation holders can decide to give their credentials to other holders. The second holder can present the credentials to a verifier. Including the information that the presented credentials are delegated and by whom may seem as the normal case. In a delegation where the second holder is not adding the information that the presented credentials were delegated may also a possibility.

### Embedding Workflow

The embedding scenario is similar to a chaining process. The difference (may) be that in embedding the claim_1/cred_1 will be presented *inside* claim_2/cred_2 and is therefore uniquely bound to 2. It may depend on the format used to represent the credentials if this bounding can be reversed.

### Attestation Workflow

In this workflow, third parties can add attestations to existing credentials, thereby enhancing their credibility. For example, after verifying a student's academic credentials, an employer could add an attestation that the student has been hired and has demonstrated certain skills on the job.

Each of these workflows offers its own set of advantages and challenges, and the choice of workflow would depend on the specific requirements of the use case at hand. By understanding and leveraging these different workflows, we can create more flexible, robust, and user-friendly systems for credential verification.

# Future Work

Looking ahead, there are several important questions that need to be addressed to further develop and refine the framework for presenting composed credentials:

1. **Revocation:** Can individual credentials or entire sets of combined credentials be revoked? Understanding how to manage revocation is crucial for maintaining trust and integrity in the system.

2. **Privacy:** Is it possible to hide certain parts of the composed credentials from the verifier? This would allow for more privacy while still enabling verification.

3. **Nested Compositions:** Can a set of composed credentials be treated as a single unit, which can then be included in another set of composed credentials? Exploring this could offer more flexibility in how credentials are presented and verified.

# Conclusion

The digital transformation of credentials is an ongoing process that presents both challenges and opportunities. This paper introduces a novel framework for the presentation of sets of related verifiable credentials, extending the traditional model of issuer, holder, and verifier to accommodate more complex and diverse use-cases. Through the concept of composed credentials, we offer a more holistic and flexible approach to credential verification, one that is adaptable to a wide range of scenarios including employment, education, and philanthropy.
Our exploration into the various workflows and processes demonstrates the versatility of this framework, providing practical solutions for real-world challenges in credential management and verification. While there are questions that require further investigation, as outlined in the Future Work section, the current framework serves as a robust foundation for the next generation of digital credentials.
By bridging the gap between stakeholders and the evolving world of digital credentials, we take a significant step towards a more secure, efficient, and user-friendly ecosystem for identity verification and trust.