Trust classification for an action's output when returned to an AI agent, derived from
the generated MJActionEntity_IActionConfiguration['outputTrustLevel'] type in @memberjunction/core-entities.
'Trusted': Internal structured data or platform-validated output (default).
'SemiTrusted': Output that may contain user-authored or partner content and should be wrapped in a data envelope.
'Untrusted': External content (web pages, search results, emails, webhooks, MCP tools) that must be wrapped in
<untrusted_external_data>, have AIDirectives stripped, and mark the agent run as tainted.
Trust classification for an action's output when returned to an AI agent, derived from the generated
MJActionEntity_IActionConfiguration['outputTrustLevel']type in@memberjunction/core-entities.'Trusted': Internal structured data or platform-validated output (default).'SemiTrusted': Output that may contain user-authored or partner content and should be wrapped in a data envelope.'Untrusted': External content (web pages, search results, emails, webhooks, MCP tools) that must be wrapped in<untrusted_external_data>, haveAIDirectivesstripped, and mark the agent run as tainted.