The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructs the agent to run npx @subframe/cli@latest sync to synchronize UI components. This is a legitimate use of the vendor's official command-line utility to maintain the local project state.
[EXTERNAL_DOWNLOADS]: The skill fetches design specifications and code structures from app.subframe.com and downloads the Subframe CLI. These network operations are directed at the vendor's infrastructure to support the intended design implementation service.
[PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it ingests untrusted design data from a remote API to generate or modify local code files.
Ingestion points: Design data is fetched from the Subframe API via the get_page_info and get_component_info MCP tools as specified in SKILL.md.
Boundary markers: There are no explicit instructions or delimiters defined to prevent the agent from following instructions that might be embedded within the external design metadata.
Capability inventory: The agent is authorized to write new files (e.g., UI components, pages) and execute shell commands through the Subframe CLI.
Sanitization: No explicit sanitization or validation of the remote design content is mentioned in the skill instructions.

develop