wayai
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFEEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTIONREMOTE_CODE_EXECUTION
Full Analysis
- [EXTERNAL_DOWNLOADS]: The instructions involve downloading the '@wayai/cli' and '@wayai/executor-sdk' from npm, as well as the third-party '@roamhq/wrtc' package for WebRTC functionality.
- [COMMAND_EXECUTION]: The skill provides procedures for setting up the developer environment, including initializing git repositories and configuring directory structures and symlinks for AI agent harness persistence.
- [INDIRECT_PROMPT_INJECTION]: The platform is designed to build agents that ingest data from external sources like 'AGENTS.md' files in repositories. This creates a surface where external content is interpolated into agent prompts without explicit sanitization mechanisms documented in the skill.
- [DYNAMIC_EXECUTION]: The skill documents the platform's 'bash' tool, which allows agents to generate and execute shell commands within a Linux-based sandbox container.
- [REMOTE_CODE_EXECUTION]: Provides templates and instructions for agents to autonomously perform repository operations, such as pushing branches to GitHub and opening pull requests.
Audit Metadata