Deployment archives should not hard-code filenames

Each function to be executed in Lambda requires its own 220MB deployment archive. Howver, the archives for different functions differs only in some filenames embedded in the lambda handler script.

By passing these filenames as event data instead, we can reuse the same lambda function for multiple tasks.

The issue names no files or tests. Start by tracing the Lambda handler script and deployment-archive flow, then verify how filenames are currently embedded and passed through event data. Done means multiple tasks can reuse one archive without hard-coded per-function filenames.

Written by the indexing model from the issue text.