Automated Post-Processing

Post-processing is the final stage of the pipeline — the automated cleanup that happens after the content is complete but before it ships. This includes formatting normalization, link validation, metadata injection, and final QA. Post-processing is not creative work. It is mechanical work that benefits from automation because humans skip it when they are tired or rushed.

The post-processing checklist should be encoded as a prompt that Claude executes against the final deliverable: "Apply the following post-processing steps to this document and return the corrected version with a log of changes made." Each step is a specific, verifiable transformation — not a judgment call.

## Post-Processing Checklist

1. [ ] All headers follow hierarchy (H1 > H2 > H3, no skips)
2. [ ] All company names match the entity glossary
3. [ ] All numbers use consistent formatting (1,000 not 1000)
4. [ ] All dates use ISO format (2026-02-22)
5. [ ] All acronyms defined on first use
6. [ ] Source list is numbered and complete
7. [ ] Word count within spec range
8. [ ] No orphaned references (cited but not listed, or listed but not cited)