Bad error messages when attempting to run invalid libraries in workflow context

We need to do a better job of capturing and emitting errors for non node compatible packages that get used inside the workflow. For example, EventTarget when trying to use the upstash redis client.

improving stacktraces (#151) is also critical to make the experience of debugging these issues better

ReferenceError while running "wrun_01K8VGKHAZ0Z96K2QZ1T1K1WD0" workflow:

evalmachine.<anonymous>:26779
var Subscriber = class extends EventTarget {
                 ^

ReferenceError: EventTarget is not defined
    at evalmachine.<anonymous>:26779:18
    at Script.runInContext (node:vm:149:12)
    at runInContext (node:vm:301:6)
    at /Users/pranaygp/github/vercel-support/v0-community-dashboard/.next/dev/server/chunks/0c935_@workflow_core_dist_cdced69a._.js:2226:158
    at /Users/pranaygp/github/vercel-support/v0-community-dashboard/.next/dev/server/chunks/0c935_@workflow_core_dist_cdced69a._.js:200:34
    at NoopContextManager.with (/Users/pranaygp/github/vercel-support/v0-community-dashboard/.next/dev/server/chunks/4b6e4_@opentelemetry_api_build_esm_9cee443b._.js:122:24)
    at ContextAPI.with (/Users/pranaygp/github/vercel-support/v0-community-dashboard/.next/dev/server/chunks/4b6e4_@opentelemetry_api_build_esm_9cee443b._.js:785:54)
    at NoopTracer.startActiveSpan (/Users/pranaygp/github/vercel-support/v0-community-dashboard/.next/dev/server/chunks/4b6e4_@opentelemetry_api_build_esm_9cee443b._.js:1962:31)
    at ProxyTracer.startActiveSpan (/Users/pranaygp/github/vercel-support/v0-community-dashboard/.next/dev/server/chunks/4b6e4_@opentelemetry_api_build_esm_9cee443b._.js:2009:24
)
    at trace (/Users/pranaygp/github/vercel-support/v0-community-dashboard/.next/dev/server/chunks/0c935_@workflow_core_dist_cdced69a._.js:198:19)
 POST /.well-known/workflow/v1/flow 200 in 182ms

x-ref thread: https://vercel.slack.com/archives/C09125LC4AX/p1761855473693419