Score the impact of renaming or deleting a field with a detailed risk assessment.

Wraps agents/field-impact-analyzer/AGENT.md. Returns a risk-scored inventory of every place a field is referenced.

Step 1 — Collect inputs

Ask the user:

1. Object API name?
   Example: Account

2. Field API name?
   Example: Industry  (or Industry__c for custom fields)

3. Target org alias (required — live-org probe is mandatory)?
   Example: prod, uat, mydevsandbox

4. Intent: rename / delete / audit?
   (Changes severity thresholds; default = audit)

5. Repo path (optional)?
   Default: force-app/main/default

If target_org_alias is missing, STOP — the agent refuses without live-org access.

Step 2 — Load the agent

Read agents/field-impact-analyzer/AGENT.md in full, plus every file under Mandatory Reads Before Starting.

Step 3 — Execute the plan

Follow the 5-step plan exactly — confirm the field exists, enumerate repo references, enumerate org references via MCP tools (list_flows_on_object, list_validation_rules, tooling_query), score risk, propose mitigation.

Step 4 — Deliver the output

Return the Output Contract block:

Summary with P0/P1/P2 + confidence
Reference inventory table
Risk breakdown
Mitigation plan (phased, no auto-patch)
Process Observations
Citations

Step 5 — Recommend follow-ups

Based on what the agent found, suggest (but do not auto-invoke):

/design-object if model-level redesign is implied
/architect-perms if FLS cleanup is surfaced
/preflight-load if the field is in upcoming loads
/detect-drift if the field appears in managed-package namespaces or unclear origins

What this command does NOT do

Does not rename, delete, or modify the field.
Does not analyze more than one field per invocation.
Does not deploy any metadata patch.