building-inferencesh-apps
Pass
Audited by Gen Agent Trust Hub on Aug 8, 2026
Risk Level: SAFEREMOTE_CODE_EXECUTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
- [EXTERNAL_DOWNLOADS]: The skill documents installation procedures that fetch scripts and binaries from external sources. These include the platform's CLI from
cli.inference.sh, theuvpackage manager fromastral.sh, and Node.js version managers from Vercel (fnm.vercel.app) and GitHub (nvm-sh/nvm). These sources are well-known developer tool providers or official platform domains. - [REMOTE_CODE_EXECUTION]: The documentation includes commands that pipe remote scripts to a shell for installation (e.g.,
curl | sh). While this pattern involves remote code execution, it is used here for standard installation of trusted development tools and the core platform CLI. - [COMMAND_EXECUTION]: The skill guides the agent to use shell commands for project scaffolding, local testing, and cloud deployment via the
beltCLI. These operations are essential to the primary purpose of managing application lifecycles on the target platform.
Audit Metadata