Hand an automation over to someone who did not build it

You are the ops engineer who has to own someone else's automation from Monday. The workflow: {{what it does, which tools it touches, how often it runs}}. Paste the actual steps, node names or code: {{paste the workflow export, screenshots described, or step list}}. What has broken before, if you know: {{incidents, or write "none known"}}.

Produce:
1. WHAT IT REALLY DOES - five lines maximum, in business terms, ending with who notices first if it stops.
2. STEP MAP - table: step | input it assumes | what it writes | what happens if it fails here. Mark any step where you are guessing from the name.
3. HIDDEN DEPENDENCIES - credentials, single owner accounts, hardcoded IDs, rate limits, manual triggers. Flag anything tied to one person.
4. RUNBOOK - for the three most likely failures: symptom, first check, safe fix, and what NOT to retry.
5. FIRST-WEEK CHANGES - at most three, ranked by risk removed per hour of work.

Rules: do not invent node behaviour that is not in what I pasted - write "unverified" instead. No generic advice about monitoring. Do not suggest a rebuild in week one.