fix(resource): idempotent DELETE — no double-deprovision (bug-bash #4/#12)

[mastermanas805]

What

A repeated DELETE /resources/:token (retry, double-click, concurrent request) re-ran SoftDeleteResource and re-issued the backend deprovision a second time. The second teardown races the first and can error against an already-gone backend — surfacing a spurious 5xx for what is logically a success.

Fix

Idempotent early-return after the team-ownership check: when the resource is already status='deleted', report success and do nothing.

if resource.Status == "deleted" {
    return c.JSON(fiber.Map{"ok": true, "already_deleted": true, "id": resource.ID.String()})
}

The first DELETE already tore the backend down; the ownership/404 posture is preserved (a resource you don't own still 404s before this check).

Coverage

Symptom:        repeated DELETE re-deprovisions backend / spurious 5xx on retry
Enumeration:    rg -F 'SoftDeleteResource' + resource.Delete handler
Sites found:    1 (single Delete handler)
Sites touched:  1
Coverage test:  TestResourceDelete_Idempotent_DoubleDelete
Live verified:  passes vs real Postgres+Redis locally; runs in CI build-and-test

First DELETE → 200; second DELETE → 200 with already_deleted=true.

🤖 Generated with Claude Code