CATEYE v5.1.0

Changelog

v5.1.0 — 2026-07-29

🔧 Integration Registry

• init_integration_registry wired into FastAPI lifespan — integrates extension registry with integration status tracking
• Import cleanup across all extension modules (aider, git, mcp, playwright, qdrant)

🔧 Infrastructure

• Import ordering and formatting consistency improvements across extension connectors and manifests

🎯 HTTP Probe Module + Widget Dashboard + Retro UI + MERLIN

• HTTP Probe Engine: protocol-agnostic probe core with HTTP/HTTPS adapter, economic scoring before requests, strategic minimal probes
• Widget Dashboard: modular widget system with 12 widget types (system health, financial, targets, findings, pipeline status)
• Retro UI Theme: retro CRT aesthetic with scanlines, phosphor glow, terminal fonts
• F1 MERLIN rebrand: MERLIN agent unified under ORION identity, logo overhaul, README refresh
• PROMOTE Stage Bridge: pipeline promotion stage for findings → evidence → report handoff
• Vision Gateway MCP: MCP server for vision model integration in the probe/evidence pipeline
• ORION Identity Docs: centralized identity and branding documentation

🧪 Testing

• 2330 tests, 100% pass rate, Ruff clean
• test_revenue_pipeline.py: 52 passed
• test_financial_hub.py: KYC, verification tracker, documents checklist
• test_target_intelligence.py: EV-based prioritizer tests
• test_ai_router.py: provider failover tests

🔧 System

• Command System Fase 1: runtime registry + dispatcher
• Session continuity audit + Command System docs
• Hermes Agent v2: EventBus integration, permission system, security layer

---

v4.5.0 — 2026-07-20

🎯 Offensive Intelligence + Revenue Pipeline

• Attack Pipeline: 6 reasoners (IDOR, SSRF, XSS, SQLi, Auth, Web3) con ProbeEngine y Evidence Composer
• Revenue Pipeline: pipeline completo Finding → Evidence → Report → Platform → Payout
• Knowledge Graph: grafo persistente con nodos/edges, query API, integración COPILOT
• Execution Runtime: simulation sandbox + task scheduler con rate limiting por plataforma
• Platform Hardening: audit log, rate limiting, CSRF, OAuth2 multi-provider

🧠 COPILOT v2

• Senior Copilot Agent: reasoning layer con planner/executor/analyzer/recommender
• Evidence Graph: for/against evidence scoring por hypothesis con persistencia
• Unified Memory: namespaces, tags, expiration, cross-session recall
• Decision Journal: decisiones con contexto + resultado + learning feedback loop
• ConfidenceScorer singleton: aprendizaje propaga al scoring en vivo

💰 Financial Expansion

• Capital Dashboard: payout summary, ROI por programa/tipo, acceptance rate, time metrics, program ranking, hot targets
• Economic Memory: ROI scoring histórico por programa y tipo de vulnerabilidad
• ATLAS Financial Intelligence: multi-agent scoring (ATLAS, MIDAS, Risk, Portfolio, F1)
• Investment Hub: gestión de inversiones con portfolio tracking

🛠️ Infrastructure

• Pre-commit hooks: Ruff (lint+format) + pytest on every commit
• Integration Center: unified view of all external integrations
• Maintenance Engine: auto-repair de servicios caídos
• Update Manager: actualización automática de pipelines
• Plugin SDK: extensión vía extensions/ auto-descubrimiento
• Hermes Doctor: diagnóstico completo del sistema

---

v4.1.0 — 2026-07-12

🎯 ORION Financial Layer

• Financial Hub: KYC Manager, Route Optimizer, Documents Checklist, Tax Notes, Payouts
• Takenos integration: USDC balance + CSV import
• Coinbase/Kraken: portfolio via HMAC auth
• Financial Intelligence: F1 multi-agent pipeline (scoring, riesgo, PnL)
• Economic dashboard con 30+ cryptos vía CoinGecko

🤖 Hermes Agent v1

• Automation Agent para ORION Platform
• Comandos: backup, status, health, logs, portfolio, prices, doctor
• EventBus integration with 3 permission levels
• Safe Mode para operaciones de mantenimiento

📚 Documentation & Audit

• Informe final del multisistema v4.1.0
• Documentación en español del README
• Night Finalization Sprint: feedback pipeline, adaptive gate, auditoría
• 3 CRITICAL bugs corregidos de seguridad y consistencia

---

v4.0.0 — 2026-07-10

🏗️ ORION Platform — Arquitectura Definitiva

• Monolito modular: EventBus para comunicación interna, arquitectura unificada
• CATEYE/AEGIS: bug bounty pipeline con Discovery → Recon → Hypothesis → Validate → Report
• ATLAS: financial hub + crypto sync + portfolio tracking
• ODYSSEY: research engine + predictive intelligence
• COPILOT: capa de inteligencia transversal con 5 niveles de autoridad
• 3 operation modes: serve (full), hunt (one-shot pipeline), SPA (frontend-only)
• AI Router: failover chain Ollama → FCC Proxy → OpenCode Free
• Target Intelligence: EV-based prioritizer con TargetPrioritizer + RewardLearner

🖥️ Frontend Overhaul

• Vue 3.5+ con Composition API, TypeScript strict mode
• Tailwind CSS v4 + ShadCN Vue components
• Mission Control: dashboard central con estado en vivo
• Command Center: Ctrl+K / Cmd+K con scopes >, /, @, #, $
• Baby Mode: interfaz simplificada para no-técnicos
• 20+ páginas: Revenue, Capital Dashboard, Financial Hub, Trading, Wallets, Intel, Knowledge Graph, Logs

🛡️ Security Hardening

• HMAC machine-id validation
• AES-256-GCM Identity Vault (~/.orion/)
• CSRF middleware con tests HTTP reales
• Rate limiting con config por endpoint
• OAuth2 multi-provider login
• Audit log rotativo (10MB, 3 backups)

---

v3.0.0 — 2026-07-08

🎯 Release Final Stable

• Marco el fin del desarrollo intensivo. CATEYE está lista para uso diario en bug bounty.

🧪 Testing

• 393 tests pasan (359 + 34 tests de seguridad), 2 xfailed, 0 fallos
• test_security.py ahora incluido en suite (34 tests, todos verdes)
• CSRF middleware verificado con tests HTTP reales via TestClient
• Rate limit middleware con tests de integración HTTP

🛠️ Correcciones de estabilidad

• Scheduler pipeline: 5 stages funcionales (DISCOVER→RECON→HYPOTHESIS→VALIDATE→REPORT)
• launch_scan() corregido (argumentos posicionales + session)
• boot_time y collect_health() bugs corregidos en api/main.py
• 3 health systems existentes pero se unificará en v3.1
• Sistema multi-agente: 8 agentes con AgentBus → EventBus bridge

🔧 Hardening (prolonged-use audit)

• FinancialSyncScheduler: sync_all() movido a asyncio.to_thread
• NotificationPoller: stop flag + shutdown hook
• Watchdog: chequea EventBus correcto (no AgentBus)
• research.py: imports y clases runner corregidos
• 14 DB indexes agregados via _migrate_indexes()
• Orphaned tasks trackeados con _background_tasks set
• WAL checkpoint agregado post-ciclo scheduler
• CorrelationEngine dedup cache limitado (10K)
• Non-blocking dispatch para BaseAgent/AgentBus
• with open() en session.py y token_service.py
• Audit log con rotación (10MB, 3 backups)

📚 Documentación

• FUNCTIONAL_SPEC.md: 988 líneas de capacidades verificadas contra código
• USER_GUIDE.md: Manual práctico en español
• DAILY_WORKFLOW.md: Rutina diaria, semanal, mensual
• RELEASE_NOTES_v3.0.0.md: Archivo eliminado en limpieza de docs redundantes
• .ai/ actualizado: AGENT_CHARTER.md, CURRENT_STATE.md, COMPLETED_FEATURES.json

📄 Licencia

• Propietaria (validación Ed25519). No MIT.

---

v2.0.0 — 2026-07-05

🚀 WalletConnect Protocol

• cores/crypto/wallet_connect.py: Conexión con wallets mobile vía WalletConnect v2
• Pairing URI generado como QR, sesión almacenada en Identity Vault
• Delegación de operaciones on-chain a EVMConnector
• POST /api/crypto/wallets/walletconnect/pair y /connect

⛓️ Crypto Expansión

• BTC connector (cores/crypto/btc.py): Blockstream.info API, satoshi→BTC, vin/vout parsing
• Solana connector (cores/crypto/solana.py): JSON-RPC mainnet, lamports→SOL, fee parsing
• Tron connector (cores/crypto/tron.py): TronGrid API, SUN→TRX, TRC20 tokens
• ERC20 tokens expandidos a 19-20 por chain (UNI, AAVE, CRV, SNX, MKR, COMP, LINK, GMX, etc.)
• 44 tests unitarios crypto con HTTP mockeado

🛡️ WithdrawalTracker Reorg-Safe

• cores/financial/withdrawal.py: +340 líneas — reescritura crypto-first
• WithdrawalEntry dataclass con tx_hash, chain, confirmations, block_number, block_hash
• DEFAULT_CONFIRMATIONS_REQUIRED por chain (ETH=12, BTC=6, SOL=30, TRX=19)
• detect_reorg(), is_reorg_safe(), auto_finalize() vía CryptoSyncManager
• ConfirmationMethod.REORG_SAFE — nuevo método de confirmación

🏦 Bank Payout Connector

• cores/financial/bank_payout.py: Plaid API + CSV import + webhooks
• PlaidProvider: link token, exchange, sync transactions, balances
• CSVImporter: parseo de CSV bancarios, auto-detección de plataformas via regex
• WebhookHandler: procesamiento de webhooks Plaid y custom
• Detección de pagos de HackerOne, Bugcrowd, Intigriti, Stripe, etc.
• 7 endpoints /api/bank-payout/*

⚡ Micro‑Functions (10)

• cores/financial/micro.py: 10 funciones compactas para el dashboard
  • quick_sync_all() — sync total con medición de tiempo
  • sync_source_now() — sync individual con delta before/after
  • get_sync_health() — salud de integraciones (success/error rate, latency)
  • trace_balance_origin() — árbol de origen del balance desde ledger
  • detect_sync_anomalies() — anomalías con severidad (negativos, stale, mismatches)
  • get_pending_actions() — acciones pendientes priorizadas
  • compute_real_exposure() — exposición financiera real (crypto + platforms + pending)
  • export_account_snapshot() — snapshot completo de cuenta
  • retry_failed_syncs() — reintento inteligente con backoff
  • get_minimal_dashboard_state() — payload ligero para UI

🖥️ Micro‑API Router

• api/routers/micro.py: 10 endpoints /api/micro/*
• 4 batch endpoints: /api/micro/batch/export|sync|delete|tag
• Entity fetch: /api/micro/entity/{type}/{id}

🎨 Micro‑Interactions Frontend (20)

• **Nuevos com...

ORION v1.0.1 — Stable Release (Hotfix)

🔧 Fixes (Root Causes 1–7)

• No QueryClientSet crash (RC#7) — QueryClientProvider moved to main.tsx, wraps entire React tree including BootScreen, LicenseGate, WelcomeWizard
• Unicode crash on Windows (RC#1) — All → (U+2192) replaced with -> in 28 files; sys.stdout.reconfigure(encoding="utf-8") on startup
• SQLite DB locked on OneDrive (RC#2) — DATABASE_URL moved to ~/AppData/Local/rastro/database/orion.db; WAL mode wrapped in contextlib.suppress
• Pipe buffer deadlock (RC#3) — All test scripts use subprocess.DEVNULL instead of subprocess.PIPE
• Log path PermissionError in Program Files (RC#5) — _setup_logging() writes to AppData instead of next to EXE when frozen on Windows
• --tray mode doesn't start backend (RC#6) — _handle_tray() now calls _run_state_machine(LaunchMode.BROWSER) instead of run_tray_only()
• console=False — Orion.exe launches without terminal window

🧪 Validation

• Playwright smoke test — 8/8 passes, zero JS errors, zero No QueryClient set, zero React crashes
• HTTP smoke test — 6 critical endpoints, 0 FAIL
• Portable test — 9/9, isolation confirmed outside repo
• Installer test — 11/11, install+run+uninstall from C:\Program Files\ORION
• Clean Windows machine test — Installed on fresh Windows with no Python/Node/Git/WSL; boot → license screen works

📦 Artifacts

• OrionInstaller.exe (41.31 MB) — one-click NSIS installer, zero runtime dependencies
• Orion.exe (16.91 MB) — standalone PyInstaller executable
• Orion-1.6.0.zip (96.71 MB) — portable distribution

Rastro v1.4.0-rc2

Changelog

v1.4.0-rc1 (2026-06-17)

• 159 tests passing (7 nuevos e2e pipeline smoke tests)
• Documentación sincronizada: README, ARCHITECTURE, ROADMAP, EVOLUTION_PLAN, MANUAL_ES, OPENCODE_PLAN
• PROJECT_STATE.md + RELEASE_CHECKLIST.md creados
• GUIA_RAPIDA.md: primer objetivo bug bounty en 9 pasos
• Fix test capacitor: test esperaban .ts, archivo es .json
• Fix prebuild.py: ya no ignora test_desktop_release.py, 16/16 checks verdes
• Fix build_windows.ps1: versión dinámica desde VERSION file
• Fix mobile/build_apk.sh: Java 17-21 check (no 22+)
• End-to-end pipeline validado: Target→Recon→Hypothesis→Investigation→Dashboard→Report
• App routes: 236 confirmadas (vs 195 documentadas previamente)
• WebSocket/WS infra: documentada como completada (ya activa desde v1.3.0)

v1.3.0 (2026-06-16)

• Pipeline Integration: Flujo unificado Mission Control → Investigation auto-creación → Pipeline/Hypothesis/Evidence/Reports scoped to investigation
• Hypothesis → Investigation: Botones "Promote to Investigation" funcionales, crean investigación con nombre generado desde hipótesis + target
• Timeline + Progreso Visual: Pipeline timeline con eventos por etapa, barra de progreso, indicador de confianza en InvestigationDetail
• Cross-navigation: Botones contextuales entre Investigation → Pipeline, Hypothesis, Evidence, Reports
• Sidebar con submenús: Navegación simplificada con submenús colapsables, sin pérdida de capacidades
• Android Scaffold: Capacitor Android inicializado, build script con Java 17+ check
• Mobile Bottom Nav: Actualizado con tabs foco mobile: Dashboard, Investigaciones, Evidencia, Reportes, Ajustes
• Empty States: Paginas principales con estados vacíos informativos y CTAs
• Fix inconsistencias: Raw fetch reemplazado por mutation hook, tipo Investigation duplicado eliminado
• Stabilization: 152 tests pasando, frontend build limpio, prebuild validation OK

v1.2.0 (2026-06-13)

• Personal Learning Engine (PLE): 7 módulos backend + 12 endpoints REST + 19 tests + UI dedicada
• Internationalization: Español default, auto-detect OS language, EN/ES completos, 22 claves PLE
• Project Governance: PROJECT_STATUS.md, TIMELINE.md, FEATURE_MATRIX.md, TECH_DEBT.md
• Project Dashboard: 7 endpoints backend + frontend con 6 tabs, health score, progress rings
• User Auth: Registration, login, refresh tokens, profile (PBKDF2-HMAC-SHA256)
• Dual DB: SQLite default + PostgreSQL via DATABASE_URL + migration script
• Consistency audit: VERSION→1.2.0, APP_VERSION lee de VERSION file, docs stale archivados
• Zero deprecation warnings: datetime.utcnow() eliminado de schemas.py y export.py
• 152 tests passing, 0 TS errors, build ~1.4s

v1.1.0 (2026-05)

• PyInstaller build para Windows + Linux
• AppImage build script
• NSIS installer para Windows
• Iconos oficiales en installer/icons/
• README premium rewrite con badges, demo flow, tree view
• UX Premium: framer-motion transitions, theme variable fixes
• 21 dead API functions eliminadas
• EVOLUTION_PLAN.md (10 semanas, 5 fases)

v1.0.0 (2026-04)

• First stable release
• Windows desktop packaging: pywebview + PyInstaller + installer
• CI/CD: test on push/PR, build Windows+Linux on tag
• Single entrypoint run.py with dev/frozen path bootstrap
• CORS hardened for desktop mode
• VERSION file as single source of truth
• Documentation: INSTALL.md (troubleshooting), PRODUCT.md, CHANGELOG.md
• Build scripts: build_frontend.sh, build_linux.sh, build_windows.ps1
• Release automation: scripts/release.py
• 111 tests passing, 0 TS errors

v0.3.0 (2026-06-13)

• AI Copilot contextual (sugerencias por ruta, Briefing + Bounty)
• Investigation Narrator layer (7 endpoints)
• Windows desktop packaging con pywebview
• PyInstaller build pipeline verificada
• Command Palette con shortcuts + recent targets
• Sidebar extraída a componente propio (-81% Layout.tsx)
• Quick Actions bar + auto-select target en MissionControl
• Builds: frontend 0 errores (897ms), tests 107/107
• Tags: v0.1-alpha → v0.2-os-ui → v0.3-ai-copilot

v0.2.0 (2026-06-12)

• Full API surface: 36 routers, 183 rutas
• Frontend OS: 22 páginas, lazy loading
• Mission Control dashboard
• System tray + notificaciones desktop
• Discovery pipeline (subfinder/katana/httpx)
• Lifespan migration + Pydantic v2 config
• CI: Makefile + scripts de build

v0.1.0 (2026-06-10)

• Initial Rastro baseline
• FastAPI + SQLAlchemy + SQLite
• React + Vite frontend shell
• Core scoring engine
• Basic recon pipeline

Rastro v1.4.0-rc1

Rastro v1.4.0-rc1 — Release Candidate

Fix crítico: Rastro.exe ya no termina silenciosamente sin WebView2.
Si el runtime WebView2 no está instalado, abre automáticamente el navegador y mantiene el backend vivo.

ZIP unificado recomendado: Rastro-1.4.0-rc1-final-unified.zip (140 MB)

Contenido

Componente	Descripción
Rastro-Windows/	Rastro.exe (16 MB, PE32) — fix startup, fallback a browser sin WebView2
Rastro-Linux/	Rastro (21 MB ELF) + run.sh
Rastro-Android/	rastro-android-debug.apk (4.2 MB)
Documentacion/	12 archivos: README, ARCHITECTURE, CHANGELOG, GUIA_RAPIDA, MANUAL_ES, RELEASE_NOTES, FINAL_PACKAGE_AUDIT, WINDOWS_RELEASE_AUDIT, DESKTOP_STARTUP_VALIDATION, etc.

Fixes en este release

• Desktop startup: _open_desktop_window() detecta cuando webview.start() retorna sin crear ventana (< 2s sin interacción del usuario) y hace fallback a browser mode automáticamente
• Desktop startup: Logs detallados antes/después de cada paso de pywebview
• Windows CI: Rebuild automático vía GitHub Actions con Python 3.12 + PyInstaller

Capacidades

• Pipeline completo: Target → Recon → Scoring → Hypothesis → Investigation → Validation → Evidence → Findings → Report
• 240 API routes, 51 routers, 28 frontend pages
• 159 tests pasando, prebuild 16/16, frontend 0 TS errors
• AI Conversacional: Ollama + OpenAI Compatible + fallback local
• 26 oportunidades activas

Limitaciones conocidas

• Windows: Sin instalador NSIS (ejecutar Rastro.exe directamente)
• Android: APK debug (no firmado para release)
• Herramientas Go: subfinder/httpx/katana/waybackurls no incluidas
• Ollama: Requerido para AI (opcional para pipeline core)
• WebView2: Si no está instalado, Rastro.exe abre el navegador en lugar de la ventana nativa (documentado)

Rastro v1.1.0-build3

Changelog

v1.0.0 (2026-06-13)

• First stable release
• Windows desktop packaging: pywebview + PyInstaller + installer
• CI/CD: test on push/PR, build Windows+Linux on tag
• Single entrypoint run.py with dev/frozen path bootstrap
• CORS hardened for desktop mode
• VERSION file as single source of truth
• Documentation: INSTALL.md (troubleshooting), PRODUCT.md, CHANGELOG.md
• Build scripts: build_frontend.sh, build_linux.sh, build_windows.ps1
• Release automation: scripts/release.py
• 111 tests passing, 0 TS errors

v0.3.0 (2026-06-13)

• AI Copilot contextual (sugerencias por ruta, Briefing + Bounty)
• Investigation Narrator layer (7 endpoints)
• Windows desktop packaging con pywebview
• PyInstaller build pipeline verificada
• Command Palette con shortcuts + recent targets
• Sidebar extraída a componente propio (-81% Layout.tsx)
• Quick Actions bar + auto-select target en MissionControl
• Builds: frontend 0 errores (897ms), tests 107/107
• Tags: v0.1-alpha → v0.2-os-ui → v0.3-ai-copilot

v0.2.0 (2026-06-12)

• Full API surface: 36 routers, 183 rutas
• Frontend OS: 22 páginas, lazy loading
• Mission Control dashboard
• System tray + notificaciones desktop
• Discovery pipeline (subfinder/katana/httpx)
• Lifespan migration + Pydantic v2 config
• CI: Makefile + scripts de build

v0.1.0 (2026-06-10)

• Initial Rastro baseline
• FastAPI + SQLAlchemy + SQLite
• React + Vite frontend shell
• Core scoring engine
• Basic recon pipeline

Rastro v1.1.0-build2

Changelog

v1.0.0 (2026-06-13)

• First stable release
• Windows desktop packaging: pywebview + PyInstaller + installer
• CI/CD: test on push/PR, build Windows+Linux on tag
• Single entrypoint run.py with dev/frozen path bootstrap
• CORS hardened for desktop mode
• VERSION file as single source of truth
• Documentation: INSTALL.md (troubleshooting), PRODUCT.md, CHANGELOG.md
• Build scripts: build_frontend.sh, build_linux.sh, build_windows.ps1
• Release automation: scripts/release.py
• 111 tests passing, 0 TS errors

v0.3.0 (2026-06-13)

• AI Copilot contextual (sugerencias por ruta, Briefing + Bounty)
• Investigation Narrator layer (7 endpoints)
• Windows desktop packaging con pywebview
• PyInstaller build pipeline verificada
• Command Palette con shortcuts + recent targets
• Sidebar extraída a componente propio (-81% Layout.tsx)
• Quick Actions bar + auto-select target en MissionControl
• Builds: frontend 0 errores (897ms), tests 107/107
• Tags: v0.1-alpha → v0.2-os-ui → v0.3-ai-copilot

v0.2.0 (2026-06-12)

• Full API surface: 36 routers, 183 rutas
• Frontend OS: 22 páginas, lazy loading
• Mission Control dashboard
• System tray + notificaciones desktop
• Discovery pipeline (subfinder/katana/httpx)
• Lifespan migration + Pydantic v2 config
• CI: Makefile + scripts de build

v0.1.0 (2026-06-10)

• Initial Rastro baseline
• FastAPI + SQLAlchemy + SQLite
• React + Vite frontend shell
• Core scoring engine
• Basic recon pipeline

Rastro v1.1.0

Changelog

v1.0.0 (2026-06-13)

• First stable release
• Windows desktop packaging: pywebview + PyInstaller + installer
• CI/CD: test on push/PR, build Windows+Linux on tag
• Single entrypoint run.py with dev/frozen path bootstrap
• CORS hardened for desktop mode
• VERSION file as single source of truth
• Documentation: INSTALL.md (troubleshooting), PRODUCT.md, CHANGELOG.md
• Build scripts: build_frontend.sh, build_linux.sh, build_windows.ps1
• Release automation: scripts/release.py
• 111 tests passing, 0 TS errors

v0.3.0 (2026-06-13)

• AI Copilot contextual (sugerencias por ruta, Briefing + Bounty)
• Investigation Narrator layer (7 endpoints)
• Windows desktop packaging con pywebview
• PyInstaller build pipeline verificada
• Command Palette con shortcuts + recent targets
• Sidebar extraída a componente propio (-81% Layout.tsx)
• Quick Actions bar + auto-select target en MissionControl
• Builds: frontend 0 errores (897ms), tests 107/107
• Tags: v0.1-alpha → v0.2-os-ui → v0.3-ai-copilot

v0.2.0 (2026-06-12)

• Full API surface: 36 routers, 183 rutas
• Frontend OS: 22 páginas, lazy loading
• Mission Control dashboard
• System tray + notificaciones desktop
• Discovery pipeline (subfinder/katana/httpx)
• Lifespan migration + Pydantic v2 config
• CI: Makefile + scripts de build

v0.1.0 (2026-06-10)

• Initial Rastro baseline
• FastAPI + SQLAlchemy + SQLite
• React + Vite frontend shell
• Core scoring engine
• Basic recon pipeline