docs: add NLM Catalog integration assessment (#1041)

Comprehensive evaluation of NLM Catalog (National Library of Medicine)
as a potential data source for biomedical journal assessment.

Key findings:
- Strong legitimacy signal for biomedical journals (MEDLINE indexing)
- Fraud detection capability (verify fake MEDLINE claims)
- Low implementation cost (1-2 developer days)
- Domain limitation: 70-80% of queries outside biomedical scope
- Recommendation: IMPLEMENT (low priority)
- Mission alignment score: 7/10

Assessment follows same format as existing evaluations (ROR, OpenAPC).
[AI-assisted]

Co-authored-by: florath-ai-assistant[bot] <Andreas.Florath@telekom.de>

Author: coding-ai-assistant[bot]

docs/future-integrations/README.md

@@ -13,11 +13,22 @@ Each file in this directory documents:
 
 ## Current Assessments
 
+- **[nlm-catalog.md](nlm-catalog.md)** - NLM Catalog (National Library of Medicine) integration assessment
+  - **Status**: Recommended / Low Priority
+  - **Reason**: Strong legitimacy signal for biomedical journals, fraud detection for fake MEDLINE claims, low implementation cost (1-2 days)
+  - **Limitation**: Domain-specific (biomedical only, 20-30% of queries), 70-80% of queries irrelevant
+  - **Implement when**: After core pattern analyzers are robust, or when biomedical user demand is clear
+
 - **[openapc.md](openapc.md)** - OpenAPC (Article Processing Charges) integration assessment
   - **Status**: Deferred / Low Priority
   - **Reason**: Limited coverage (5-15% of queries), weak signal for predatory detection, cost data ≠ quality indicator
   - **Reconsider if**: Scope expands to include cost transparency, user demand, or for research use cases
 
+- **[ror.md](ror.md)** - ROR (Research Organization Registry) integration assessment
+  - **Status**: Not Recommended
+  - **Reason**: Fundamental mismatch (institutions ≠ journals), catastrophically low coverage (<10%), extremely weak signal, massive complexity (2,437 LOC, 7 tables)
+  - **Related**: PR #1034 fully implemented and tested ROR integration before rejection
+
 ## Adding New Assessments
 
 When evaluating a new potential integration: