The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted input data from the Input Model JSON structure, which is then interpolated into output blocks like TUI_RENDER, COMPONENT_SPEC, and PENCIL_BATCH_DESIGN.
Ingestion points: The Input Model JSON object, specifically the props, modelValue, and state.error fields in SKILL.md.
Boundary markers: The instructions do not define delimiters or specific "ignore instructions" warnings to separate untrusted data from the agent's logic.
Capability inventory: The skill generates design specs and pseudo-code (PENCIL_BATCH_DESIGN) intended for consumption by the Pencil MCP tool; it does not directly execute shell commands, perform network operations, or write to the filesystem.
Sanitization: There is no evidence of sanitization, validation, or escaping of the input data before it is placed into the output templates, potentially allowing malicious strings to influence the agent or downstream tools.

tui-color