Repeatability
High
The task is a one-time code generation job with a fixed structure: fetch, diff, log, patch, test. Every instance of this request has the same shape, making it highly automatable.
Ambiguity Tolerance
High
Success criteria are explicit — reconcile every 30 minutes, handle three named edge cases, include unit tests. An agent can verify its own output against these criteria without human interpretation.
Data & Tool Availability
Medium
Shopify GraphQL and REST API patterns are well-documented in training data, so the agent can write correct-looking code. However, actual API credentials, the fulfillment center's specific endpoint schema, and the local cache format are not provided, so the output will need configuration before it runs.
Error Cost
Medium
Buggy reconciliation logic could POST incorrect stock corrections to the fulfillment center, causing oversells or stockouts. The risk is real but bounded — the script is a corrective layer, not a payment processor, and bad updates can be rolled back with a re-sync.
Human Judgment Required
Low
The reconciliation rules are deterministic: mismatches get flagged and corrected. No taste, ethics, or relationship context is needed — just solid Python and API integration logic.