I’ll help write an RFC, design doc, code, etc.

Reply to this note

Please Login to reply.

Discussion

Might be good to have a fallback in the spec, something that emails abuse@domain.com perhaps.

Here's a high-level outline for such a protocol:

1. Introduction

- Purpose of the protocol

- Scope of the protocol

- Definitions and terminology

2. Protocol Overview

- Explanation of the protocol flow

- Roles and responsibilities of the involved parties

3. Identification and Verification

- Process for identifying and verifying the reporting entity

- Use of digital certificates or other secure methods for verification

4. CSAM Reporting Mechanism

- Detailed description of how to report CSAM

- Required information in the report (e.g., URL of the image, timestamp, any additional context)

- Use of secure and encrypted communication channels for reporting

5. Image Host Response

- Process for the image host to verify the report

- Steps for the image host to take upon verification (e.g., immediate takedown of the image, reporting to law enforcement)

6. Fallback Mechanism

- Process for sending an email to security@domain or abuse@domain if the primary reporting mechanism fails

- Adherence to ISO 29147 guidelines for vulnerability disclosure

7. Audit and Compliance

- Regular audits to ensure compliance with the protocol

- Penalties for non-compliance

8. Privacy and Security Considerations

- Measures to protect the privacy and security of the reporting entity

- Measures to prevent misuse of the protocol

9. Protocol Updates and Versioning

- Process for updating the protocol and maintaining different versions

10. Conclusion

- Summary of the protocol

- Contact information for queries related to the protocol

This is a high-level outline and each section would need to be expanded upon with detailed procedures and technical specifications. It's also important to involve legal, technical, and child protection experts in the development of the protocol to ensure it is effective, lawful, and ethical.

Shouldn’t be hard to put out a call for more help.

Did you just use chatgpt 😆

Nah, if I was going to cheat I’d use Claude2 instead. But I have two RFC under development right now so I just copied and pasted from those outlines and made relevant changes. And I missed removing a bit about vulnerability disclosure. 😢🫣