The Agent Skills Directory

[COMMAND_EXECUTION]: The skill uses the Bash tool to create project directory structures through commands such as mkdir -p plugin-{name}/.... While this is a standard functional requirement for scaffolding, it introduces a surface where malformed user input for the plugin name (e.g., containing shell metacharacters) could potentially lead to command injection if not handled by the agent's underlying safety filters.\n- [PROMPT_INJECTION]: The skill processes user-defined capabilities and names to generate code, which represents a surface for indirect prompt injection.\n
Ingestion points: User-provided inputs during the requirements analysis phase, including plugin names, descriptions, and action details.\n
Boundary markers: No specific delimiters or instructions to ignore embedded commands within user input are provided in the workflow.\n
Capability inventory: The skill uses Bash, Write, Edit, Read, Grep, and Glob tools to manage the file system and project files.\n
Sanitization: The skill's instructions do not define validation or sanitization steps for user-supplied strings before they are interpolated into commands or file templates.

plugin-builder