The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill documentation provides command templates that use placeholders such as {project}, {api-service}, and <your-server>. These are intended to be replaced with user-supplied values and then executed in a shell environment.
Ingestion points: Placeholders in SKILL.md (e.g., {project}, {api-service}, <your-server>, <registry-host>).
Boundary markers: None specified in the instructions to separate code from data.
Capability inventory: Execution of local bash scripts, docker CLI commands (both local and remote via contexts), ssh commands, and curl requests.
Sanitization: No explicit sanitization or validation of the placeholder values is mentioned, which could allow for command injection if a user provides input containing shell metacharacters (e.g., ;, &, |, `).
[COMMAND_EXECUTION]: The skill's core functionality involves executing multiple local bash scripts located in scripts/server/ and running commands on remote servers via SSH. While this is the intended purpose, it grants the agent broad execution capabilities over the local and remote environments.

deploy-remote