Add conservative identifier readability check

[MrC8559]

Summary

Implements the names/identifiers half of #61 with a deliberately conservative AST check.

• splits snake_case and CamelCase declaration names into tokens;
• checks function, async-function, and class names only;
• flags clearly compressed vowel-stripped tokens such as chk / cfg;
• flags multiple short consonant-only tokens such as fp_sz;
• preserves common technical initialisms and unknown all-uppercase acronym chunks;
• treats y as a vowel to reduce false positives such as sync;
• registers the check in the existing CHECKS registry.

The heuristic intentionally prefers false negatives over false positives. It is not a dictionary, general English checker, or claim of standards conformance.

Tests

Added focused coverage for:

• snake_case / CamelCase / acronym tokenization;
• clear compressed-name positives;
• functions, async functions, and classes;
• common technical initialisms, digits, acronym chunks, and readable-name negatives;
• non-declaration AST nodes;
• registry wiring.

Validation

Final upstream gate run #201 passed completely on commit e148ff86a39306789c14ce953549f01a947bf4bc:

• Ubuntu gate: passed
• macOS gate: passed
• Windows gate: passed
• dependency review: passed

The gate includes the repository's audit, security, pytest/100%-coverage, Pyright, Ruff format/lint, Pylint, and complexity checks. Earlier CI feedback was used to remove regex, satisfy the repository's docstring/style rules, simplify the tokenizer below the complexity cap, and restore full coverage/format compliance.

Scope

This does not touch the Vale/prose half of #61.

[rxdt]

MrC8559, thank you for this. You took on a hard part of the issue and kept it practical. The rule is clear, and the tests make its limits easy to see. I appreciate the care you put into it.

[MrC8559]

Thanks!
Glad I could help.