ВозможностиДокументацияЦеныPlaygroundFAQ

Roadmap

FlowLink эволюционирует из MCP gateway в полноценную AI-native SecOps & Operations Platform. Вот наш путь — что доставлено, что в работе, куда двигаемся.

5

фаз доставлено

1

в разработке

2

запланировано

Фаза 1: Core Security

✅ Доставлено
  • MCP Gateway (inline, WSS/HTTP)
  • Shield Engine (pattern matching, risk scoring)
  • Policy Engine (per-agent, per-tool, per-service)
  • Approval Queue (Telegram, Web UI, API)
  • Audit Log (каждое действие записано)
  • Command History (exit codes, durations, shield results)
  • Rate Limiting (per-agent, hot-reloadable)
  • Authentication (OAuth VK/Yandex/GitHub, email, 2FA)

Фаза 2: Zero-Trust & Secrets

✅ Доставлено
  • Secret Injection (env vars, files, vault pull)
  • Per-Org Encryption (X25519 + AES-256-GCM)
  • External Vault Integration (HashiCorp Vault)
  • Zero-Trust API (key setup, verification, rotation)
  • E2EE Agent Communication

Фаза 3: Infrastructure & Observability

✅ Доставлено
  • Infrastructure Map (семантический граф сервисов)
  • Service Discovery (auto-detect через агентов)
  • Service Catalog (ownership, SLA, health, risk)
  • Agent Health Monitoring (heartbeat + auto-recovery)
  • SIEM Integration (Alertmanager, Generic webhooks)
  • Billing & Subscriptions (4-tier pricing)

Фаза 4: Forensics & AI Ops

✅ Доставлено
  • Incident Timeline (blast radius + anomaly detection)
  • Agent Reconstruction (scenario replay)
  • Compliance Reports (security audit, policy compliance)
  • Context Snapshots (point-in-time state + diff)
  • AI Ops Assistant (natural language queries)
  • Efficiency Insights (agent ROI, time saved)
  • Change Management (approval + rollback)
  • Telegram Bot (webhook mode)

Фаза 5: GitOps & DevOps

✅ Доставлено
  • ServerGuard (eBPF мониторинг файлов, Docker events, canary tokens)
  • Drift Detection (nginx, systemd, SSH configs)
  • Auto-Backup перед destructive operations
  • GitOps Bridge (agent↔relay sync)
  • K8s Operator (CRD + admission webhook)
  • Restore API + CLI commands

Фаза 6: Intelligence & Scale

🔄 В разработке
  • 🔄Pattern Learning (auto-suggest политик из поведения агентов)
  • 🔄Anomaly ML (behavioral baseline, drift detection)
  • 🔄Cost Attribution (per-agent, per-service cost tracking)
  • 🔄Budget Alerts (spending limits per agent/org)
  • 🔄Multi-tenant RBAC (fine-grained роли и permissions)

Фаза 7: Platform & Ecosystem

📋 Запланировано
  • 📋TypeScript SDK (программное управление платформой)
  • 📋Python SDK (автоматизация и скриптинг)
  • 📋Prometheus Metrics Exporter
  • 📋Grafana Dashboard Templates
  • 📋Terraform Provider (Infrastructure as Code)
  • 📋Kubernetes Helm Chart
  • 📋Plugin System (custom Shield rules, integrations)
  • 📋Marketplace (community-contributed policies и patterns)

Фаза 8: Advanced Security

📋 Запланировано
  • 📋Prompt Injection Detection (анализ tool outputs)
  • 📋Tool Poisoning Prevention (валидация MCP server signatures)
  • 📋Data Loss Prevention (DLP patterns в tool responses)
  • 📋Session Recording (полный terminal replay)
  • 📋Threat Intelligence Feed (known malicious patterns)
  • 📋Compliance Automation (SOC2/ISO27001/ФСТЭК)

Влияние и выравнивание

Наш roadmap основан на:

  • OWASP GenAI MCP Security Guidetool poisoning, least-privilege, runtime guardrails
  • Enterprise AI Security Standardszero-trust, audit compliance, incident response
  • ФСТЭК / ГОСТ Р 57580российский compliance для government и enterprise
  • Community feedbackGitHub Issues
Есть идея? Откройтеissue на GitHubили напишите нам на[email protected]или вTelegram