fix((extra..)time

fix((extra..)

fix((extra..)

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

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

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

feat(compliance): Gold Standard Deployment - Strict JSON, Stability Fixes, & Full Coverage Tests

GOLD STANDARD: JSONResponse Bypass + Msg>=2 Callback + Handshake Fix

WINNER-TIER: Enforce Strict GUVI Response Schema (Status + Reply Only)

WINNER-TIER: Single Source of Truth + Predictive Finalization + Safe Guards

FINAL: Clean logging, lifecycle guarantee, latency fix

CRITICAL FIX: GUVI Lifecycle Overhaul (Handshake loop, empty msg, callback trigger)

HOTFIX: Syntax Error in guvi_handler.py (Duplicate argument removed)

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

fix(handshake): Implement processStatus init response to resolve GUVI retry loop

fix(schema): Make message field optional to robustly handle GUVI initialization pings

feat(behavior): Implement human-like typing delays (Req 7 compliance)

fix(routes): Remove redundant GUVI callback to prevent empty intel overwrite

feat(compliance): Enable Mandatory GUVI Final Result Callback

fix(registry): Disable JSON_SCHEMA for Llama-3.3 on Groq to prevent 400 errors

fix: Optimize LLM client latency (fail-fast 400s, swift key rotation)

fix: Resolve NoneType attribute error in persona engine and orchestrator

fix: Key rotation now tries rotation FIRST for multi-account keys

fix: Dashboard confidence display and GUVI response format

fix: Dashboard now correctly shows confidence score from API

fix: Make telemetry beacon flexible to accept short field names

fix: Restore FULL GUVI response format matching Jan 28 working version

debug: Add detailed GUVI request logging to diagnose e917 error

fix: Return minimal GUVI response per Section 8 spec - Only status + reply fields to fix e917 INVALID_REQUEST_BODY

docs: Restore official Winner-Tier requirements with full analytics support

docs: Add requirements documentation and restore requirements.txt

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

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

fix: explicit persistence of scam keywords to memory

feat: ensure critical intel visibility in agent notes

fix: syntax error and restore email/crypto extractors

feat: integrate full threat intel pipeline

feat: upgrade to SOC-grade intelligence extraction

feat: fast-track banking urgency scenarios

fix: memory leak lazy cleanup

fix: prevent double-normalization of intelligence in callback

feat: enable fast-track callback for one-shot tests

fix: imports in callback_client.py

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

🏆 ANALYTICS BONUS: Added Scam Confidence, Risk Level, and Timeline for Final GUVI Scoring

🛡️ SOC-GRADE: Implemented Verhoeff Aadhaar validation, hardened NPCI UPI patterns, and context-aware bank account scanning

🚀 EXTRACTION UPGRADE: Robust handling for 9-digit bank accounts and refined tactical pattern matching

🚀 PRODUCTION READY: Hyper-robust schema hardening, safe history injection, and enhanced debug logging for 100% compliant GUVI evaluation

🚀 DEBUG: Loosened GUVI schemas and added 422 request body logging for deep diagnostics

🚀 DEFINITIVE COMPLIANCE: Doubled message count, restricted intel keys to 5-key spec, and synchronous callback for platform stability