Classifier gaps: 7 new regex patterns in schemata v0.3.2

[alltheseas]

Context

schemata v0.3.2 introduced 9 new regex patterns that fall back to regex in the classifier. 2 are already covered by #38 (hex_alternation, base64_2pad). The remaining 7 need new native ops.

Unclassified patterns

From the classifier-gate CI failure on #39:

#	Pattern	Proposed op	Description
1	^[a-z][a-z0-9]*$	identifier	lowercase identifier
2	^[A-Z][a-zA-Z0-9]*$	identifier	PascalCase identifier
3	^[a-z][a-z0-9-]*$	identifier	lowercase kebab identifier
4	^!?[a-z][a-z0-9]*$	identifier	optional ! prefix + lowercase
5	^!?[0-9]+$	optional_prefix_digits	optional ! prefix + digits
6	^[a-z_]+( [a-z_]+)*$	space_separated_charset	space-separated lowercase+underscore tokens
7	^[A-Za-z][A-Za-z0-9+.-]*://	uri_scheme	URI scheme prefix (no end anchor)

Approach

Patterns 1-4 share a common shape: ^[optional_prefix][first_char_class][rest_char_class]*$. A single identifier op with configurable fields could cover all four.

Pattern 5 is similar but digit-only body.

Pattern 6 extends the existing space_separated_tokens concept.

Pattern 7 is a URI scheme prefix check (unanchored end).

Blocked by

• feat: add hex_alternation, base64_2pad, identifier, space_separated_charset, uri_scheme native ops #38 (hex_alternation + base64_2pad must land first)

Related

• chore: bump schemata to v0.3.2 #39 (automated bump PR, CI failing due to these gaps)