12 posts tagged with "Inteligência Artificial"

Refactored the AI Pentest creation screen for a clearer, more guided setup experience.

• Reorganized the form so configuring a new AI Pentest is more intuitive.
• A smoother setup flow that makes it easier to launch well-scoped autonomous tests.

👉 Access the Conviso Docs to experience these updates.

Expanded the AI Pentest flow with new white-box capabilities and documentation.

• Azure integration for white-box pentest scenarios.
• Privacy by design: source code is never sent to the LLM a key differentiator of the approach.
• Expanded official documentation for the AI Pentest workflow.

Watch: Pentest Autônomo: Como Ganhar Escala em Segurança Ofensiva com IA.

Added configuration levels and scheduling to the AI Pentest, giving teams more control over how and when automated tests run.

• Define configuration levels to tailor the depth and scope of each AI Pentest.
• Set automated schedules, moving away from manual triggers toward continuous offensive validation.

Watch: Exploração de Vulnerabilidades: o que é e como a IA faz diferente.

Improved the Conviso MCP Server and published new governance documentation.

• Broader compatibility: the MCP now works across all Claude products — Desktop, Cowork/Projects, and CLI.
• Improved MCP documentation for easier setup and usage.
• New Skills documentation.
• Documented data retention policy.

👉 Access the Conviso Platform to experience these updates.

Introduced a new policy framework specifically for automated pentests, mirroring the Security Gate logic but optimized for offensive security workflows.

• Define specific goals for an asset, set request limits to ensure safety during execution, and define automated schedules.
• Moves teams away from manual testing triggers, ensuring assets are continuously validated against offensive scenarios.

Introduced the AI Autonomous Pentest, an autonomous offensive security engine that combines Large Language Models with professional security tools to reduce manual toil and scale testing.

• Tool Orchestration: automatically triggers and correlates data from tools and findings.
• Offensive Reasoning: decides the best attack path, identifies vulnerability chains, and executes controlled Proof of Concepts (PoCs).
• Attack Chain Visualization: a real-time interactive graph shows the AI's reasoning from reconnaissance to successful exploit validation.

Watch the launch webinar: AI Autonomous Pentest.

Officially launched the Conviso MCP (Model Context Protocol) Server, available on the Anthropic MCP-Market, allowing external AI models to interact securely with your Conviso Platform data.

• Query assets, vulnerability status, and project updates directly within the chat interface.
• Brings Conviso's security intelligence into the developer's AI workspace, enabling faster remediation suggestions and architectural reviews without leaving the LLM environment.

Release date: April 6th, 2026

Key Benefits​

• AI Autonomous Pentest;
• Conviso MCP Server;
• Full-Snapshot DAST Reporting;

Introduction​

Release 4.24 expanded autonomous security operations, AI ecosystem connectivity, and the completeness of DAST reporting.

---

New Feature

AI Autonomous Pentest​

• Introduced an autonomous offensive-security engine that orchestrates tools and validates attack paths.
• Added attack-chain visualization to provide transparency from reconnaissance through exploit validation.

Learn more in the AI Pentest documentation.

---

New Feature

Conviso MCP Server​

• Made the Conviso MCP Server available for secure AI-assisted access to Platform context.
• Enables approved AI tools to use vulnerability, asset, and project context in remediation and review workflows.

---

Improvements

Full-Snapshot DAST Reporting​

• Changed DAST reports from a delta-only view to a full snapshot of the application's security posture for each execution.
• Provides an auditable, self-contained record of vulnerabilities identified at scan time.

The Threat Modeling module now functions as a living record of your application's security design — a persistent artifact that evolves throughout the software lifecycle instead of a project that "closes".

• Persistent Architecture-as-Code: use visual diagrams as the primary driver to update your living security model without starting from scratch.
• Automated Requirement Sync: instantly translate architectural changes into actionable security controls.
• Historical Traceability: native versioning provides a complete audit trail of every design shift, ensuring full compliance and governance.

Release date: March 23rd, 2026

Key Benefits​

• Continuous Living Threat Modeling;
• Defect Tracker Recent Deliveries;

Introduction​

Release 4.23 made Secure by Design workflows more continuous and improved visibility into the delivery of findings to engineering backlogs.

---

New Feature

Continuous Living Threat Modeling​

• Introduced living, versioned threat-modeling artifacts driven by architecture diagrams.
• Keeps security requirements synchronized with architectural changes and preserves their evolution as an auditable history.

---

New Feature

Defect Tracker Recent Deliveries​

• Added visibility into recent synchronization events for defect trackers such as Jira, ClickUp, and ServiceNow.
• Helps teams troubleshoot delivery issues with detailed context and status information.

Expanded the AI Agents operations across the platform.

• Introduced Auto-Fix, which can be triggered directly from vulnerability context.
• Expanded False Positive review support to DAST issues, increasing consistency across scanning sources.
• Improved AI Agent chat usability with Start new conversation to reset context and return to quick actions.
• Multiple stability and quality improvements across agent execution and recurring runs.

Release date: October 24th, 2025

Key Benefits​

• Threat Modeling Agent (AI-Powered);
• Performance Optimization;
• Enhanced PDF Reports;
• Executive Reports;
• Multilingual Platform & Reports;
• AI Agent – Contextual Fix Suggestions;
• AI Agent – Vulnerability Lookup by ID;
• Retest Project Improvements;
• Filter System Redesign;
• DAST – False Positive Reduction (ML-Based);

Introduction​

This release brings multiple AI-powered capabilities, language flexibility, and major usability upgrades.
Highlights include multilingual reports, contextual AI fixes, enhanced retest workflows, and expanded threat modeling intelligence.
These updates aim to make the Conviso Platform faster, more intuitive, and even more powerful for global security teams.

New Feature

Threat Modeling Agent (AI-Powered)​

• New AI-driven threat modeling agent available in the Threat Modeling module.
• The AI agent supports this input formats for modeling:
  • Images: PNG, JPEG, and PDF architectural diagrams.
  • User Stories: Markdown (.md) or plain text (.txt) files.
• Automatically recognizes and models assets from both visual and textual sources.
• The agent parses the file, identifies components, and generates CAPEC-based security requirements automatically.
• Examples include disabling GraphQL introspection, enforcing rate limiting, and query cost analysis.
• Helps standardize and accelerate threat modeling with minimal manual input.

---

Improvements

Performance Optimization​

• Reduced initial load time from ~9 seconds to under 100ms.
• The platform now delivers faster page rendering and improved responsiveness.

---

New Feature

Enhanced PDF Reports​

• Technical and executive reports now support inline image rendering.
• Key updates include:
  • Display of assets associated with each project.
  • Inclusion of the full project timeline and individual requirement timelines.
  • Inline images for vulnerabilities and requirements.
  • Non-image attachments automatically bundled in a .zip file.
• Result: richer, clearer, and more contextualized reports for both technical and managerial use.

---

New Feature

Executive Reports​

• New Executive Report type designed for decision makers.
• Provides concise, result-oriented summaries without deep technical detail.
• Keeps customizable text fields for contextual information.
• Removes complexity filters and focuses on key project outcomes.
• Simplifies reporting processes and ensures consistent communication across stakeholders.

---

New Feature

Multilingual Platform & Reports​

• Added language selection in the user profile, allowing interface translation between English and Portuguese.
• Reports (Technical, Executive, and DAST) can now be generated in multiple languages.
• Improves accessibility for international users and enhances comfort for non-English speakers.
• Users can report translation feedback directly to the team for quick improvement.

---

New Feature

AI Agent – Contextual Fix Suggestions​

• The AppSec AI Agent now provides step-by-step fix recommendations directly from the vulnerability view.
• New “agent-ai” icon under Solution opens a chat pre-filled with the vulnerability context.
• The agent explains and guides remediation automatically.
• Reduces friction and accelerates vulnerability remediation processes.

---

New Feature

AI Agent – Vulnerability Lookup by ID​

• The AI chat now supports commands such as “how to fix vulnerability #12345”.
• Allows direct retrieval of vulnerability details from the MCP Server by ID.
• Enables smarter contextual assistance inside the AppSec AI Agent chat.
• Improves demos and client visibility during vulnerability review sessions.

---

Improvements

Retest Project Improvements​

• Added new functionality to associate vulnerabilities more easily with retest projects.
• Vulnerabilities can be linked by:
  • Asset, Project, Scan ID, or Vulnerability ID.
• Simplifies retest creation and improves overall efficiency in validation workflows.

---

Improvements

Filter System Redesign​

• Introduced redesigned filter management system across key modules.
• Filters are now persistent, shareable, and URL-based, maintaining state across sessions.
• Improves navigation and usability across Assets, Projects, Vulnerabilities, and Scans.

---

Improvements

DAST – False Positive Reduction (ML-Based)​

• The DAST scanner now uses machine learning to detect and exclude generic error pages automatically.
• Reduces the number of false positives reported in dynamic application tests.
• Delivers cleaner, more accurate vulnerability reports.

---

👉 Access the Conviso Platform to explore these updates in action.