The Agent Skills Directory

[COMMAND_EXECUTION]: The skill relies on the execution of the commet CLI to handle authentication, organization management, and project linking operations. These commands are executed locally as part of the standard developer workflow.
[EXTERNAL_DOWNLOADS]: Through the commet create command, the skill fetches application templates from GitHub repositories. These downloads are directed to the vendor's official resources (commet-labs) for project initialization.
[DATA_EXFILTRATION]: The skill involves managing sensitive data locally, specifically storing credentials in ~/.commet/auth.json and writing API keys to .env files. This behavior is standard for CLI authentication and project configuration.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface by downloading project templates and generating TypeScript definition files (.commet/types.d.ts) from remote dashboard configurations.
Ingestion points: Project templates from GitHub and type definitions pulled via the commet pull command.
Boundary markers: None identified in the provided instructions for isolating generated or downloaded code.
Capability inventory: The skill executes shell commands via the CLI and performs file system writes to store configuration and environment variables.
Sanitization: No explicit sanitization or integrity verification of the remote data is described in the skill instructions.

commet-cli