This document aims to explain the process of identifying a security issue and the steps for managing a security issue.
It is commonly stated that a security issue is an issue that:
Given that WASI is a set of Capability-based APIs, all unauthorized actions are not supposed to happen. Most of the above security concerns can be alleviated. What remains for us is to ensure that the execution of Wasm modules is secure. In other words, do not compromise the sandbox. Unless it is explicitly disabled beforehand.
Thus, we share most of the criteria for judging security issues with the Bytecode Alliance.
[!NOTE] keep updating this document as the project evolves.
Follow the same guidelines as other projects within the Bytecode Alliance.
Before reporting an issue, particularly one related to crashing, consult the cheat sheet, Report a security vulnerability if it qualifies.
Upon receiving an issue, thoroughly review the cheat sheet to assess and Report a security vulnerability if the issue is indeed a security vulnerability.
Once a security issue is confirmed, please refer to the runbook for the subsequent steps to take.