The Agent Skills Directory

[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection because it processes untrusted user descriptions to guide its UI enhancement and code generation workflows.
Ingestion points: User prompts are ingested and processed by the enhance-prompt and stitch-design workflows in SKILL.md.
Boundary markers: There are no specific delimiters or instructions to prevent the model from obeying instructions embedded within the user-provided prompt.
Capability inventory: The skill utilizes Bash, Write, and WebFetch tools to execute code generation, file organization, and network requests based on those prompts.
Sanitization: No input validation or filtering of the user-provided content is documented or required before processing.
[EXTERNAL_DOWNLOADS]: The skill documentation includes instructions for installing tools and components from remote repositories.
Downloads logic: The skill instructs users to install packages and plugins via npx and marketplace commands from the google-labs-code organization and the author's akillness/jeo-skills repository.

stitch-skills