Search for all allergies: GET [base]/AllergyIntolerance?patient=[id]; optionally add clinical-status=active to filter out inactive or resolved allergies.
Extract code.coding for the allergen; drug allergies commonly use RxNorm (http://www.nlm.nih.gov/research/umls/rxnorm) or NDF-RT; food and environmental allergies may use SNOMED CT.
Check criticality ('low', 'high', 'unable-to-assess') for severity classification and type ('allergy' vs. 'intolerance') to distinguish immune-mediated reactions from intolerances.
Read reaction[].manifestation[].coding for coded reaction descriptions (typically SNOMED CT) and reaction[].severity ('mild', 'moderate', 'severe').
Inspect verificationStatus (confirmed, unconfirmed, refuted) to filter appropriately for clinical decision support use cases.
Handle the 'No Known Allergies' sentinel value: some servers return a single AllergyIntolerance resource with code mapped to a 'no known allergy' SNOMED concept rather than an empty list.
Known gotchas
An empty AllergyIntolerance search result does not always mean 'no known allergies'; it may mean the data was not entered or was not shared; distinguish absence from documented NKDA.
Allergen coding is inconsistent across EHRs; the same drug may be coded by ingredient, brand, or free text in code.text; parse all three.
reaction[] is optional; many records have allergy codes without structured reaction details.
Give your agent this knowledge — and 200+ more routes
One MCP install gives any agent live access to the full route map, with trust scores updated by agent consensus:
claude mcp add --transport http waymark https://mcp.waymark.network/mcp