2026

Improvements

• Apply security update

New Features

• Git mode — Manage test specs as CSV files in a GitHub repository and review them in the same branch and pull request as your code. Edit, commit, create working branches, and open PRs right from the KATATAN UI
• Record the source (file, branch, commit) of each Git-mode test report, and show a notice with a GitHub diff link when the spec is updated after the report was created
• Add `katatan report create` (create a test report from a local CSV) and `katatan report export` (export a report to CSV) in `@katatan/cli` 0.2.0
• Add MCP tools `create_git_report` and `export_test_report` (existing MCP access tokens must be edited to allow them)

Improvements

• Redesign the test spec table editor (CSV4TEST), with auto-save in DB mode and on-screen resolution of concurrent edit conflicts
• Show No and test subject on public share pages as they were when the report was created
• Export rows for deleted test cases in report CSV exports using their contents at report creation time

New Features

• Add `@katatan/cli`, a command-line client with the same functionality as the MCP API, so AI agents and CI/CD pipelines don't need to load tool schemas into context
• Add delete operations for test specs, test cases, test reports, labels, dictionary entries, and test case images, available via both the MCP API and the CLI

New Features

• Add domain verification to Vulnerability Check

Improvements

• Apply security updates to dependencies to address known vulnerabilities

New Features

• Add Vulnerability Check — Automatically run 14 security checks against project-registered URLs and generate vulnerability reports

Improvements

• Apply security update

New Features

• Workspace management — Manage resources per project with member invitations and RBAC (role-based access control)
• Test specs & test cases — AI agents automatically generate and record test cases via the API
• Test reports — Record execution results in a structured format with three statuses: pass / fail / skip
• Label management — Attach custom key-value labels to test cases and test reports, configurable via API
• MCP support — Model Context Protocol tools for seamlessly operating Katatan from AI agents
• REST API — Public endpoints for workspaces, test specs, test cases, and test reports