Changelog
v1.1.1
New Features
- Download integrity verification — SHA256 checksum verification for downloaded files. On failure, prompts manual download.
- Download mirror support — Mirror selector in settings with automatic fallback and skip-verification option.
- AI thinking display — AI conversations now show thinking process, message queue, stop/retry, and error recovery.
- Terminal context menu & search — Right-click context menu and search in terminal. MySQL TLS support, password echo.
- Edge reveal strip — Edge hot zone component for sidebar navigation.
- System status dashboard — Startup status registry with settings page section and toast notifications for migration failures.
Bug Fixes
- Fixed DeepSeek and other OpenAI-compatible API returning 400 errors.
- Fixed CI golangci-lint v2 formatting and staticcheck issues.
- Removed redundant
nolint gosecdirectives. - Fixed UpdateSection prettier formatting.
CI/CD
- Generate
release-info.jsonon release for mirror update checks.
Other
- README rewrite highlighting use cases and security auditing.
Full Changelog: v1.1.0...v1.1.1
v1.1.0
New Features
- Multi-provider AI configuration — Database-backed provider management with full CRUD API. Added native Anthropic Messages API support alongside OpenAI-compatible providers.
- AI Agent enhancements — Dynamic system prompts, sub-agent support, parallel tool execution, and batch command processing.
- Unified approval system — Merged AI and opsctl approval dialogs into a single workflow. Supports batch confirmation, Grant mechanism with editable command patterns, and remember mode.
- AI model intelligence — Model parameter presets with auto-fill on selection, context window auto-compression, and Anthropic prompt caching optimization.
- Multi-asset permission requests —
request_permissiontool now supports multi-assetitemsparameter, consistent withopsctl grant. - opsctl improvements — Redis
-nflag for database selection;create/update assetsupports--iconparameter.
UI/UX Improvements
- AI setup wizard redesign — New opsctl plugin banner, improved provider selection cards, extracted shared
AIProviderFormcomponent. - AI chat UI refactoring — Chat header, role labels, solid assistant bubbles, independent approval block rendering.
- Terminal Aesthetic theme — New visual style for the application interface.
- AI settings enhancement — Model list fetching from API, model selector with search, token parameter configuration.
- Approval block improvements — Differentiated rendering for single/batch/grant types; only pending approvals interrupt message flow.
- ToolBlock & AgentBlock refinements — Optimized spacing, line height, running state indicator, error state icons.
- GFM Markdown support — AI chat now renders GitHub Flavored Markdown (tables, strikethrough, task lists).
- Database & Redis UX — Improved frontend user experience for query modules.
Bug Fixes
- Fixed Windows subprocess causing a brief black window flash on launch.
- Fixed new SSH asset defaulting to port 6379 instead of 22.
- Fixed
SSHPool.Close()panic on double-call. - Fixed opsctl
exec_sql/exec_redis/cpaudit logs missing decision and decision source.
Backend Improvements
- User denial handling — Strong stop instructions when user denies commands; system prompt includes denial guidance to prevent AI from attempting workarounds.
- Concurrency safety — Approval callbacks now receive conversation ID via context instead of instance variable.
- Error messages in AI module standardized to English for better LLM comprehension.
Other
- Added GPLv3 License.
- Updated README links to opskat.github.io.
- Added i18n translations for approval buttons and batch operations.
Full Changelog: v1.0.2...v1.1.0
v1.0.2
New Features
- opsctl batch command — Batch execution of SSH commands, file transfers, and database/Redis operations across multiple assets.
- Batch approval dialog — Approve multiple operations at once from the desktop app.
- Plugin system refactor — Migrated
skill/directory toplugin/, supporting multi-platform plugin registration (Claude Code, Codex, Gemini CLI). - Build info module — Embed commit ID into binary for version traceability.
Testing & Quality
- Added 16 frontend test files covering stores, hooks, components, and utilities.
- Added opsctl batch command tests and GitHub backup service tests.
- Improved golangci-lint error handling: replaced
//nolintwith proper logger-based error handling foros.Remove/os.RemoveAll.
CI/CD
- Release/Nightly build optimization.
- Integrated frontend Vitest tests into CI.
Other Improvements
- Frontend component accessibility attributes.
- Added i18n translations for batch operations.
- GitHub backup service refactoring.
- Auto-update service fixes.
Full Changelog: v1.0.1...v1.0.2
v1.0.0
OpsKat is a desktop application for server asset management and remote operation and maintenance. It integrates an AI intelligent assistant and supports SSH terminal, database queries, Redis operations, and file transfer.
Main Functions
- Asset Management — Grouping and management of server, database, and Redis assets, supporting encrypted credential storage.
- SSH Terminal — Multi-tab terminal, supporting split-screen, jump server chaining, port forwarding, and SOCKS proxy.
- Database Query — MySQL/PostgreSQL query editor, supporting SSH tunnel connections.
- Redis Operations — Redis command execution, supporting SSH tunnel connections.
- SFTP File Management — Remote file browsing and transfer.
- AI Smart Assistant — Supports multiple backends including OpenAI compatible API, Claude CLI, and Codex CLI, enabling operation and maintenance through natural language.
- Security Policy — Three types of policies: SSH commands, SQL statements, and Redis commands, supporting policy group inheritance and pre-approval authorization.
- Audit Log — All operations are automatically recorded and traceable.
- opsctl CLI — Independent command-line tool supporting asset management, remote execution, file transfer, and other operations.
Full Changelog: v1.0.0