The Agent Skills Directory

[SAFE]: The skill defines a strict, case-sensitive specification for LMX (Loops Markup Language). It explicitly forbids the use of unknown tags or attributes, which prevents the accidental or malicious generation of non-standard or dangerous markup.
[SAFE]: A mandatory output checklist requires the agent to escape XML-sensitive characters (&, <, ") in both text content and attribute values. This is a critical security measure to prevent injection attacks within generated email bodies.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process external, untrusted source material (such as existing HTML, MJML, or screenshots) to migrate them into LMX. The skill manages this risk effectively through the following:
Ingestion points: Untrusted data enters the context through user-provided source material and migration requests.
Boundary markers: The instructions explicitly state that source material cannot override the skill's internal safety and specification rules.
Capability inventory: The skill's primary capability is generating LMX markup and creating visual references via the imagegen tool.
Sanitization: The skill mandates strict XML escaping and tag validation as a post-processing step for all generated content.
[SAFE]: All referenced tools and workflows (imagegen, loops-api, loops-cli) are aligned with the skill's stated purpose of email creation and management within the Loops platform.

loops-lmx