The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill processes external data (metadata and text) from the Mobbin API, which represents a theoretical attack surface for indirect prompt injection. However, the skill includes explicit security instructions for the agent to treat all returned content as untrusted reference data and specifically warns against executing commands or following links found within that data.
Ingestion points: Data returned by search_screens, search_flows, and search_sections tools from api.mobbin.com.
Boundary markers: The 'Evidence handling' section explicitly instructs the agent to ignore any commands or links embedded in the reference data.
Capability inventory: The skill records local research data to research/<topic>/ and interacts with a companion design skill.
Sanitization: The skill relies on behavioral instructions to prevent the agent from acting on potentially malicious external content.
[SAFE]: The connection to https://api.mobbin.com/mcp is consistent with the skill's stated purpose of integrating with Mobbin's official UI research service. The connection commands provided are standard for the Codex environment.

mobbin-usage