The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user-provided data, such as existing interface code and design system files, to generate critiques and new design patterns. It instructs the agent to save these patterns to a local system file (.interface-design/system.md), creating a vulnerability surface where malicious instructions embedded in the analyzed code could influence the agent's actions or corrupt the design system.
Ingestion points: The skill ingests existing interface code, design requirements, and screenshots from the user, as specified in SKILL.md and references/design-workflow.md.
Boundary markers: Absent. The instructions do not define specific delimiters or instructions to ignore embedded prompts within the user-provided code snippets.
Capability inventory: The skill possesses the capability to write and update patterns in the .interface-design/system.md file, as described in references/design-workflow.md and references/validation.md.
Sanitization: Absent. There is no mention of filtering, escaping, or validating the content of the ingested code before it is processed or saved to the file system.

interface-design