Standardize digest algorithm label to BLAKE-512 — revert BLAKE2b-512 occurrences in NFH-007 and beckn.yaml

The RequestDigest.digest schema in api/v2.0.0/beckn.yaml (~line 3478) uses the prefix BLAKE-512= in its validation pattern, while [NFH-007 (Authentication and Trust)](https://github.com/beckn/protocol-specifications-v2/blob/main/docs/Authentication_and_Trust.md) and the rest of beckn.yaml consistently use BLAKE2b-512=. Rather than updating the one outlier to match, this issue proposes going the other way: BLAKE-512 is the correct/intended label, and the BLAKE2b-512 occurrences elsewhere are the drift that needs correcting.

Scope of the change:

docs/Authentication_and_Trust.md (NFH-007) — replace all occurrences of BLAKE2b-512 with BLAKE-512, including:
Definitions (§ "Body digest") and the algorithm summary table
Signing-string construction pseudocode and worked examples (digest = "BLAKE2b-512=" + Base64(...))
Step 5/6 of the callback and request verification procedures (§7, §8)
Conformance requirement CON-004-04 ("The body digest MUST be computed using BLAKE2b-512...")
The References section entry for RFC 7693
api/v2.0.0/beckn.yaml — replace all remaining BLAKE2b-512 occurrences (the Authorization/Signature header descriptions, the Ack and callback schema descriptions, and inline examples) so they match the corrected RequestDigest.digest pattern, leaving exactly one label in use across the spec.
Open question for the working group: RFC 7693 specifically defines the BLAKE2 family (BLAKE2b/BLAKE2s), not the original BLAKE algorithm. If the intent is to keep computing digests the BLAKE2b way but simply refer to it as "BLAKE-512" for brevity, the References section needs to make that explicit (e.g. "informally referred to as BLAKE-512 in this document; the underlying primitive is BLAKE2b-512 per RFC 7693") so implementers don't mistakenly implement the original, different BLAKE-512 algorithm instead.