The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill provides instructions for the agent to interpolate content from the conversation (the "brief") directly into shell command arguments for the devspace CLI.
Ingestion points: The <brief> and <follow-up brief> parameters in the devspace agents run and devspace agents continue commands described in SKILL.md are intended to be populated using input from the conversation.
Boundary markers: The documentation does not provide specific delimiters or instructions to the agent to ignore potentially malicious instructions embedded within the brief data.
Capability inventory: The skill utilizes the agent's shell or process execution tool to run the devspace CLI, which can spawn further processes and perform complex tasks.
Sanitization: There are no instructions for sanitizing or validating the contents of the brief before it is passed to the command line, creating a surface where untrusted data could influence subagent behavior or attempt command injection if the underlying shell tool does not handle quoting robustly.
[COMMAND_EXECUTION]: The core functionality of the skill is centered on executing local shell commands via the devspace CLI to manage agent sessions and project tasks. While this is the intended purpose, it provides the agent with the capability to execute arbitrary sub-processes.

subagents