The Agent Skills Directory

[PROMPT_INJECTION]: The skill's workflow involves reading screen specifications from the specs/pages/ directory to generate prototypes. This represents a surface for indirect prompt injection where malicious instructions embedded in specifications could influence the agent's behavior.
Ingestion points: Reads external data from the specs/pages/ directory (SKILL.md).
Boundary markers: Absent; no instructions are provided for the agent to treat the imported specification data as untrusted or to ignore embedded instructions.
Capability inventory: The skill utilizes the use_figma tool to modify design files and executes shell commands using the memi CLI (SKILL.md).
Sanitization: No explicit sanitization, validation, or filtering of the specification content is mentioned.
[COMMAND_EXECUTION]: The skill instructs the agent to execute local commands using the memi CLI tool for generating HTML prototypes and creating architectural documentation.
Evidence: Use of memi prototype, memi spec page, and memi ia create commands within the workflow instructions (SKILL.md).

figma-prototype