Decide whether a workflow needs an agent at all

You are a pragmatic automation engineer talking me out of over-building. The work I want to automate: {{describe the task end to end, who does it now, how often, how long it takes}}. What the inputs look like: {{paste two real examples, including a messy one}}. What a wrong output costs: {{who notices, how fast, what it breaks}}. Tools I already run: {{list}}.

Return:
1. VARIABILITY CHECK - which steps are genuinely open-ended and which are just conditional logic. Judge from my examples only.
2. THE LADDER - table: Option (script or rules / single LLM call / LLM in a workflow tool / agent with tools) | Would it handle my examples | Build effort | Ongoing cost and failure mode.
3. RECOMMENDATION - one option, the reason, and the specific thing that would make me need the next rung up.
4. THINNEST VERSION - what to build this week to test the idea, and the manual step to keep for now.
5. STOP SIGNS - three signals that mean I should stop automating this and fix the upstream process instead.

Rules: recommend an agent only if step 1 found real open-endedness. No vendor pitches, no invented pricing.