The Agent Skills Directory

[COMMAND_EXECUTION]: The skill operates by executing shell commands via npx @cardelli/ambit, flyctl, and tailscale to provision and manage cloud infrastructure.
[EXTERNAL_DOWNLOADS]: The deploy --template functionality fetches application source code and configuration files directly from GitHub repositories.
[REMOTE_CODE_EXECUTION]: Templates downloaded from GitHub are deployed as active services on Fly.io, which involves the execution of the template's container definitions and build logic.
[CREDENTIALS_UNSAFE]: The skill requires users to manage and provide sensitive credentials, specifically Tailscale API tokens and Fly.io authentication states, which are necessary for its core functionality.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface through its template ingestion mechanism.
Ingestion points: Untrusted application data enters the environment through the --template <owner/repo> argument used in deployment commands.
Boundary markers: There are no documented boundary markers or instructions to ignore potential commands embedded within the external template content.
Capability inventory: The skill possesses significant capabilities, including application deployment, secret management, and network policy modification.
Sanitization: While the documentation mentions a 'pre-flight scan' for dangerous settings, the specific sanitization or validation logic is not detailed.

ambit-cli