π Metadata & Policy
How entities describe capabilities and superiors constrain them
What is Metadata?
Every entity in a federation describes its capabilities through metadata[1] β structured key-value pairs organized by Entity Type Identifier. The metadata tells other entities what protocols are supported, which endpoints are available, and how to interact.
Metadata by Entity Type
Β· issuer
Β· authorization_endpoint
Β· token_endpoint
Β· userinfo_endpoint
Β· jwks_uri
Β· scopes_supported
Β· response_types_supported
Β· grant_types_supported
Β· subject_types_supported
Β· id_token_signing_alg_values_supported
Policy Operators
Superiors constrain their subordinates' metadata using policy operators[2] in the metadata_policy claim of Subordinate Statements. There are 7 operators:
value
one_of
subset_of
superset_of
add
default
essential
Policy Cascade β Watch It Work
Metadata policies are applied in order from the Trust Anchor down through each Intermediate to produce the final resolved metadata.[3]
Step 1 β Leaf's Raw Metadata
Real-World Analogy
Think of job requirements: an employee lists their skills (metadata), the department head limits them to 3 relevant skills (intermediate policy), and the CEO adds a company-wide rule "nothing outside business hours" (TA policy). The resolved result is what the employee actually does β shaped by every level above.