quest-tooling
Pass
Audited by Gen Agent Trust Hub on Nov 12, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADSCOMMAND_EXECUTIONPERSISTENCEINDIRECT_PROMPT_INJECTION
Full Analysis
- [REMOTE_CODE_EXECUTION]: The skill includes instructions to download a shell script from Meta's official developer portal (
https://developers.meta.com/horizon/install-cli/) and pipe it directly to a shell for installation. This is the documented native installation method for the Meta VR CLI. - [EXTERNAL_DOWNLOADS]: The skill utilizes
npxto download and execute themetavrpackage from the NPM registry and facilitates the installation of managed developer tools (e.g.,perfetto,renderdoc,xrsim) via themetavr tools installcommand. These resources originate from established and official platforms. - [INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from external sources, including documentation searches (
metavr docs search) and third-party samples. It explicitly implements defensive instructions, warning the agent that these inputs are untrusted and must not be allowed to override host configuration, authorization protocols, or security boundaries. - [COMMAND_EXECUTION]: The skill provides a detailed command map for the
metavrCLI, allowing the agent to perform actions such as application installation, file manipulation on connected devices, and UI automation. These capabilities are consistent with the skill's stated purpose of developer tooling. - [PERSISTENCE]: The skill contains procedures for registering MCP (Model Context Protocol) servers in the agent host's configuration. This allows the tools to be persistently available to the agent across different sessions, which is standard functionality for MCP-integrated skills.
Audit Metadata