Changelog

What's new in HoundShield

Every improvement, shipped fast.

v2.7.0July 2026

Control Reference Library & Partner Kit

  • FeatureAll 110 NIST 800-171 controls published as a free reference — official text, plain English, remediation, assessor evidence, and an honest AI-relevance verdict per control
  • FeatureRPO/MSP partner kit: co-branded CMMC AI Risk Assessment at published wholesale terms
  • FeatureConsole assessment and gap views now deep-link every control to its full guide
  • ImprovementScanner latency is now a CI-enforced contract: 2,000 cold scans per build must keep p99 under 10ms (currently 0.57ms)
v2.6.0July 2026

Customer Status Intelligence & Personalized Brain AI

  • FeatureAfter-login status panel: where you stand, your single next step, and your highest-impact gaps with fixes
  • FeatureBrain AI answers “where do I stand?” from your own data — consent-gated, off by default, revocable
  • FeatureOrder confirmation for the $499 CMMC AI Risk Assessment Report — reference, amount, report-due date, live status
  • ImprovementPrivacy by design: assessment data stays in your browser; no customer data is ever sent to a commercial LLM
v2.5.0March 2026

CMMC Level 2 Full Mapping

  • FeatureComplete 110-control CMMC Level 2 mapping with practice-level detail
  • FeatureSPRS score calculator with real-time gap weighting
  • Featureassessor-reviewable report export (PDF + JSON)
  • ImprovementHoundShield assessment engine refactored for sub-200ms scoring
v2.4.0February 2026

AI Agent Workspace

  • FeatureReAct reasoning loop with chain-of-thought tracing
  • Feature12 integrated tools for compliance automation
  • Feature13 AI model support via OpenRouter gateway
  • ImprovementAgent templates for common compliance workflows
v2.3.0January 2026

Stripe Integration & Pricing

  • FeatureStripe checkout for the $499 one-time AI Risk Assessment Report
  • FeaturePromo code support at checkout
  • FixWebhook idempotency for duplicate Stripe events
v2.2.0December 2025

Command Center V2

  • FeatureUnified dashboard with real-time threat feed
  • FeatureCompliance metrics panel with trend sparklines
  • ImprovementPipeline view for scan lifecycle tracking
  • FixWebSocket reconnection reliability on flaky networks
v2.1.0November 2025

OAuth & Multi-Provider Auth

  • FeatureGoogle, GitHub, and Microsoft SSO providers
  • FeatureMagic link passwordless authentication
  • ImprovementSession management with secure HTTP-only cookies
v2.0.0October 2025

Complete Platform Rebuild

  • FeatureFull rewrite on Next.js 15, React 19, Tailwind CSS
  • FeatureDark-mode-first design system
  • FeatureZero-copy stream scanning engine
  • ImprovementSub-50ms P99 scan latency
Roadmap

Coming next

Later this year

Agent run grouping

Collapse the hundreds of prompts in one autonomous agent run into a single reviewable unit, with a named human owner.

today412 separate events · sender "a process" · no owner
nextRun #1184 · Claude Code · 412 prompts · 3 blocked · owner: your engineer
Later this year

Per-agent policy

Different rules for people and for agents. Hold an unattended run for review on patterns a human would only be warned about.

todayOne policy for every caller, human or not
nextHumans: warn on PII · Agents: quarantine on PII, no exceptions
Later this year

Browser extension

Catch the AI tools that never touch the proxy — anything used straight in a browser tab.

todayGateway covers API and SDK traffic
nextGateway + browser tab: chat.openai.com, claude.ai, Gemini, Copilot

Ship with confidence

Start with the free in-browser scan — no signup required.

Get started