The Agent Skills Directory

[REMOTE_CODE_EXECUTION]: The skill repeatedly instructs the user to install a CLI using curl -fsSL https://cli.inference.sh | sh. This pattern executes a remote shell script directly from the internet without prior inspection, which is a high-risk operation that could be exploited if the remote source is compromised.
[EXTERNAL_DOWNLOADS]: The skill facilitates downloading and executing binaries from dist.inference.sh. While the documentation describes verification steps (SHA-256 and Sigstore), the primary installation flow bypasses manual verification in favor of the automated shell pipe.
[CREDENTIALS_UNSAFE]: The references/authentication.md file encourages users to store sensitive API keys in environment variables (export INFSH_API_KEY=your-api-key). While this is a common practice, the skill's instructions specifically target the management of these secrets for use with the external service.
[COMMAND_EXECUTION]: The skill makes extensive use of the Bash tool to interact with the local filesystem and the belt CLI. This includes commands that automatically upload local files (e.g., ~/.ssh or .env files if misconfigured by the user) to a remote cloud provider for processing.
[DATA_EXFILTRATION]: The CLI is designed to automatically upload local file paths provided in JSON inputs to the inference.sh cloud. This capability, while functional for the skill's purpose, creates a path for sensitive local data to be sent to external servers.

agent-tools