The Agent Skills Directory

[COMMAND_EXECUTION]: The skill defines multiple workflows and launch configurations that utilize powerful command-line utilities including kubectl, docker, git, npm, psql, and pg_dump. These tools are employed for standard developer operations such as deploying services, managing containers, and performing database backups.
[EXTERNAL_DOWNLOADS]: The provided templates include instructions for cloning repositories from GitHub and installing project dependencies using npm install. These actions are part of the intended environment setup functionality and target well-known services.
[INDIRECT_PROMPT_INJECTION]: The skill's examples describe the agent processing user requests to generate terminal configurations and shortcuts, which creates a potential attack surface for indirect prompt injection.
Ingestion points: Natural language user requests for custom terminal setups or functional extensions (found in SKILL.md examples).
Boundary markers: No explicit delimiters are specified in the templates to separate user-provided parameters from executable commands.
Capability inventory: The workflows utilize high-privilege tools such as kubectl, docker, and git across various project directories.
Sanitization: The skill does not demonstrate explicit sanitization or validation of variables passed into the command templates.

warp