Replace non-deterministic fallback IDs with explicit skip logic in EAC-CPF indexing

[Copilot]

The traject configuration used Time.now.to_i and rand() to generate fallback IDs for creator records lacking standard identifiers. This broke idempotent indexing—the same record would receive different IDs across runs, creating duplicates instead of updates.

Changes

• Remove non-deterministic fallback: Replace timestamp/random ID generation with context.skip!()
• Add explicit error logging: Surface missing ID issues with clear context about what failed (no recordId, invalid filename pattern, missing entity data)

Rationale

Records reach the fallback only if:

1. No <control>/<recordId> in XML
2. Filename doesn't match creator_{type}_{id}.xml pattern
3. No entity type/name to derive ID from

This indicates a data pipeline failure, not missing data. The normal flow (task_agent() → deterministic filenames) provides reliable IDs. Skipping malformed records preserves index integrity and surfaces issues for investigation rather than masking them with synthetic IDs.

# Before: Non-deterministic
fallback_id = "creator_unknown_#{Time.now.to_i}_#{rand(10000)}"
accumulator << fallback_id

# After: Explicit skip
context.logger.error("Cannot generate valid ID for record - skipping indexing. Source: #{source_file}")
context.skip!("Missing required ID data")

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.