Optimize: Cache compiled regex patterns in constraint matcher

[mediumWellness]

Summary

Adds LRU caching to regex compilation in the constraint matcher engine, eliminating repeated RegExp constructor calls on hot paths.

Changes

• Add \PATTERN_CACHE\ Map at module level with max 1000 entries
• Extract flag parsing into separate \�uildPatternWithFlags()\ helper
• Check cache before compiling each pattern
• Implement simple LRU eviction on cache overflow

Performance Impact

Estimated: 30-50ms saved per 100 evals with repeated constraint patterns

Rationale:

• Each \\ constraint evaluation calls \�uildPattern()\
• Regex compilation is O(pattern length) — significant for validators like (?i)^[a-z]+$\
• Common validation patterns (email, phone, zipcode) repeat across test suites
• Cache key is deterministic (pattern string) — guarantees correctness

Testing

• ✅ All 157 matcher tests pass
• ✅ Backwards compatible (no API changes)
• ✅ No existing behavior changes

Files Changed

• \webmcp-evals/src/matcher.ts\ (+34 lines) — regex cache implementation

Related

Part of 10-item performance audit (ranked #1 by impact/effort)

[google-cla]

Thanks for your pull request! It looks like this may be your first contribution to a Google open source project. Before we can look at your pull request, you'll need to sign a Contributor License Agreement (CLA).

View this failed invocation of the CLA check for more information.

For the most up to date status, view the checks section at the bottom of the pull request.

[swissspidy]

Can we close this in favor of #390, or why the duplicate PRs? Also note the lack of CLA signing.