docs: add CLAUDE.md behavioral guide and PR_REVIEW_GUIDE.md

[ankita10119]

🔧 Changes

Add CLAUDE.md and PR_REVIEW_GUIDE.md to encode development and review standards directly into the repo.

CLAUDE.md - behavioral guide for working in this repo (for both humans and AI agents):

• Feature development protocol: step-by-step process for implementing handlers or API specs, including pre-code questions to answer, the four required handler methods (getType, calcChanges, processChanges, validate), test requirements, and format verification
• Pre-submit gate: four questions that must be answered before marking work complete, plus required local checks
• deploy-cli invariants: non-negotiable constraints for every handler/context parser change (both formats, dry-run, keyword replacement, write-only/read-only field stripping, identifier matching, ordering, schema)
• EA / entitlement-gated feature rules: how to document untested code paths and API assumptions when a feature requires special entitlement to test end-to-end
• Traps: high-signal patterns from real bugs in this codebase

PR_REVIEW_GUIDE.md - living checklist for thorough PR reviews, covering 13 areas: context, correctness, edge cases, backward compatibility, safety, tests, docs, code quality, performance, security, CI, deploy-cli specifics, and EA/entitlement-gated scenarios. Includes final sign-off questions and lessons learned from real reviews.

📚 References

N/A

🔬 Testing

No functional code changes - documentation only. Verify by reading both files and checking that the deploy-cli specific claims match the codebase (handler methods, keyword replacement patterns, stripCreateFields/stripUpdateFields, @order() usage).

📝 Checklist

• All new/changed/fixed functionality is covered by tests (or N/A)
• I have added documentation for all new/changed functionality (or N/A)