fix: Made honeypot act naive/dumb - remove security awareness and formal punctuation

fix: Reduced typo intensity, added self-correction, improved multi-turn handling

feat: Enhanced human typo injection system - realistic keyboard errors based on age/profession

Shorter realistic responses: mood-based length, max 80 tokens like real SMS/chat

Fix persona lock: embed selected_persona_key in get_persona return

Clean up responses: fix fraud word replacement, reduce repetitive fillers and stuttering

Fix: Increase max_tokens to 180, add Hindi-to-English cleanup for English mode

Fix: Strengthen English language enforcement in prompt

Fix: Humanize responses - increase tokens to 120, persist language override across turns

fix((extra..)time

fix: Make time_utils language-aware for English/Hindi responses

fix: Stronger language enforcement - explicit forbidden Hindi words for English mode

fix: Strict language enforcement in prompts

fix: Remove 'scammer' word leak and improve human-likeness

Fix language adaptation, URL extraction, persona consistency

CRITICAL: Fix 413/400 errors and reduce API calls by 50%

fix((extra..)time

fix(guvi): Resolve Handshake Loop - Strict Schema (Data vs Reply) & Data Population

Final critical GUVI fixes: callback threshold, crash fix, zero-sleep mode

fix: Resolve NoneType attribute error in persona engine and orchestrator

Deployment Ready: Fixed scam detection low confidence, added production audit report, optimized throttles

feat: Platinum Grade Hardening & Zero-Trust Compliance Fixes (Audit v4.2)

feat: Final Hackathon Submission - Autonomous Adaptive Agent V2.0

✨ Added self-correction phrases to personas for GUVI compliance

🔥 FINAL POLISH: Dynamic Personas & System Verified

🔧 Fixed Persona Identity Consistency (Dynamic Name Injection)

🔥 Added Decoy Victim Profiles (Realistic Financial Data)

🔥 Enterprise Agentic Scam Honeypot v2.0 - India AI Buildathon 2025