Discover, compose, run, publish, and save Adobe Firefly workflows through the run-workflow MCP server.
93
96%
Does it follow best practices?
Impact
83%
1.25xAverage score across 3 eval scenarios
High
Do not use without reviewing
{
"context": "Tests that naming a known featured workflow (e.g. retargeting) routes through get_featured_workflow's fast path instead of composing a graph from scratch via compose_workflow. Skipping this check wastes a compose step and risks producing a worse, non-template graph when a vetted one already exists.",
"type": "weighted_checklist",
"checklist": [
{
"name": "Calls get_featured_workflow first",
"description": "The agent called get_featured_workflow (with a query reflecting 'retargeting') before any call to compose_workflow.",
"max_score": 35
},
{
"name": "No premature compose_workflow",
"description": "The agent did not call compose_workflow when a named featured workflow was requested and a match was available, skipping the from-scratch graph-building path.",
"max_score": 25
},
{
"name": "Uploads local assets correctly",
"description": "The agent called upload_asset for the local product photos and used the returned presigned URLs verbatim (no truncation or reconstruction) in the submission.",
"max_score": 15
},
{
"name": "Uses session_id submission",
"description": "When a single prepared match was returned, the agent submitted via run_workflow_submit using session_id and useSessionDefaults rather than re-sending the full actions/connections graph.",
"max_score": 15
},
{
"name": "Reports output or status",
"description": "The agent reported back batch/run status or output URLs to the user rather than ending the turn silently after submission.",
"max_score": 10
}
]
}