Refactor new-device QR flow and streamline bypass init:
- napcat-shell: stop verbose logging and removed check of enableAllBypasses return value; just invoke native enableAllBypasses when not disabled by env.
- backend (QQLogin): simplify extraction of tokens from jumpUrl (use sig and uin-token), return an error if missing, and send oidbRequest directly (removed nested try/catch and regex fallback).
- frontend (new_device_verify): accept result.str_url without requiring bytes_token and pass an empty string to polling when bytes_token is absent.
- frontend (password_login): change render order to show captcha modal before new-device verification UI.
- frontend (qq_manager): normalize GetNewDeviceQRCode response — derive bytes_token from str_url's str_url query param (base64) when bytes_token is missing, and preserve extra status/error fields in the returned object.
These changes improve robustness when OIDB responses omit bytes_token, reduce noisy logs, and ensure the UI and polling still function.
Introduce multi-step QQ password login support (captcha and new-device verification) and related OIDB QR handling.
- Change login signature fields in NodeIKernelLoginService to binary (Uint8Array) and add unusualDeviceCheckSig.
- Update shell base to handle additional result codes (captcha required, new-device, abnormal-device), set login status on success, and register three callbacks: captcha, new-device, and password flows. Use TextEncoder for encoding ticket/randstr/sid and newDevicePullQrCodeSig.
- Extend backend WebUiDataRuntime (types and runtime) with set/request methods for captcha and new-device login calls and adjust LoginRuntime types to return richer metadata (needCaptcha, proofWaterUrl, needNewDevice, jumpUrl, newDevicePullQrCodeSig).
- Add backend API handlers: CaptchaLogin, NewDeviceLogin, GetNewDeviceQRCode and PollNewDeviceQR; add oidbRequest helper using https to query oidb.tim.qq.com for QR generation and polling.
- Wire new handlers into QQLogin router and return structured success responses when further steps are required.
- Add frontend components and pages for captcha and new-device verification (new files: 1.html, new_device_verify.tsx, tencent_captcha.tsx) and update existing frontend controllers/pages to integrate the new flows.
- Improve error logging and user-facing messages for the new flows.
This change enables handling of password-login scenarios requiring captcha or device attestation and provides endpoints to obtain and poll OIDB QR codes for new-device verification.
Introduce support for plugin icons across backend and frontend. Updates include:
- napcat-onebot: add optional `icon` field to PluginPackageJson.
- Backend (api/Plugin, PluginStore, router): add handlers/utilities to locate and serve plugin icons (`GetPluginIconHandler`, getPluginIconUrl, findPluginIconPath) and wire the route `/api/Plugin/Icon/:pluginId`.
- Cache logic: implement `cachePluginIcon` to fetch GitHub user avatars and store as `data/icon.png` when package.json lacks an icon; invoked after plugin install (regular and SSE flows).
- Frontend: add `icon` to PluginItem, prefer backend-provided icon URL in plugin card (via new getPluginIconUrl util that appends webui_token query param), and add the util to handle token-based image requests.
- Plugin store UI: add a Random category (shuffled), client-side pagination, and reset page on tab/search changes.
These changes let the UI display plugin icons (falling back to author/avatar or Vercel avatars) and cache icons for better UX, while handling auth by passing the token as a query parameter for img src requests.
Set bypass defaults to disabled and simplify loading: napcat.json default bypass flags changed to false and code now reads bypass options without merging a prior "all enabled" default. Removed the progressive bypass-disable logic and related environment variable usage, and added a log when Napi2NativeLoader enables bypasses. Web UI/backend adjustments: default NapCat config is now generated from the AJV schema; the bypass settings UI defaults to false, adds an o3HookMode toggle, and submits o3HookMode as 0/1. UX fixes: extension tabs made horizontally scrollable with fixed tab sizing, and plugin uninstall flow updated to a single confirmation dialog with an optional checkbox to remove plugin config. Overall changes aim to use safer defaults, simplify crash/restart behavior, and improve configuration and UI clarity.
Introduce a new 'process' bypass option and remove the old 'maps' key, updating all related schemas, types, defaults and validation. Adjusted napcat.json ordering and NapcatConfig defaults, updated BypassOptions schema and TypeScript interfaces, backend validation keys, loader/default parsing logic, and the web UI form mappings/labels to reflect the new field set and ordering.
Introduce granular "bypass" configuration to control Napi2Native bypass features and expose it in the WebUI.
Key changes:
- Add bypass defaults to packages/napcat-core/external/napcat.json and BypassOptionsSchema in napcat-core helper config.
- Extend Napi2NativeLoader types: enableAllBypasses now accepts BypassOptions.
- Framework & Shell: load napcat.json (via json5), pass parsed bypass options to native loader, and log the applied config. Add json5 dependency.
- Shell: implement loadBypassConfig with a crash-recovery override (NAPCAT_BYPASS_DISABLE_LEVEL) and add master<->worker IPC (login-success) plus progressive bypass-disable strategy to mitigate repeated crashes before login.
- WebUI backend: add GET/Set endpoints for NapCat config (NapCatConfigRouter) with validation and JSON5-aware defaults.
- WebUI frontend: add BypassConfig page, types, and controller methods to get/set bypass config.
- Update package.json to include json5 and update pnpm lockfile; native binaries (.node / ffmpeg.dll) also updated.
This enables operators to tune bypass behavior per-installation and to have an in-UI control for toggling anti-detection features; it also adds progressive fallback behavior to help recover from crashes caused by bypasses.
Add end-to-end support for reading/writing Linux machine-info and computing GUIDs.
Backend: - Introduce MachineInfoUtils (TS) for machine-info path lookup, ROT13 serialization, read/write/delete, backups, and MD5-based GUID computation. - Add a Python utility (guid.py) for CLI inspection, encode/decode, dump, and GUID computation. - Extend QQLogin API with new handlers: GetPlatformInfo, GetLinuxMAC, SetLinuxMAC, GetLinuxMachineId, ComputeLinuxGUID, GetLinuxMachineInfoBackups, CreateLinuxMachineInfoBackup, RestoreLinuxMachineInfoBackup, ResetLinuxDeviceID. Handlers include automatic backup behavior and error handling.
Router: Register new QQLogin routes for platform info and Linux machine-info operations.
Frontend: - Enhance guid_manager UI to detect platform and provide Linux-specific workflow (display machine-id, show/edit MAC, preview computed GUID via MD5, backup/restore/delete machine-info, and restart actions). - Add client-side MD5 (crypto-js) usage and new QQManager API methods to call the new backend endpoints.
This change enables cross-platform GUID management (Windows and Linux), includes CLI tooling for low-level inspection, and adds frontend workflows for Linux device-id management.
Introduce tools and UI to read, write, backup and restore QQ Registry20 GUIDs using Windows DPAPI. Adds a Python CLI (guid_tool.py) for local Registry20 operations and a new TypeScript package napcat-dpapi with native bindings for DPAPI. Implements Registry20Utils in the webui-backend to protect/unprotect Registry20, plus backup/restore/delete helpers. Exposes new backend API handlers and routes (get/set GUID, backups, restore, reset, restart) and integrates frontend GUIDManager component and qq_manager controller methods. Propagates QQ data path via WebUiDataRuntime (setter/getter) and wires it up in framework/shell; updates Vite alias and package.json to include the new dpapi workspace. Includes native addon binaries for win32 x64/arm64 and basic tsconfig/readme/license for the new package.
Implement password-based QQ login across the stack: add a PasswordLogin React component, integrate it into the QQ login page, and add a frontend controller method to call a new /QQLogin/PasswordLogin API. On the backend, add QQPasswordLoginHandler, router entry, and WebUiDataRuntime hooks (setPasswordLoginCall / requestPasswordLogin) plus a default handler. Register a password login callback in the shell (base.ts) that calls the kernel login service, handles common error cases and falls back to QR code when needed. Update types to include onPasswordLoginRequested and adjust NodeIKernelLoginService method signatures (including passwordLogin return type changed to Promise<QuickLoginResult>) and minor formatting fixes.
Introduce support for plugin API routes that do not require WebUI authentication. Updates include:
- napcat-onebot: Add apiNoAuth route storage and helpers (apiNoAuth, getNoAuth, postNoAuth, putNoAuth, deleteNoAuth), hasApiNoAuthRoutes, buildApiNoAuthRouter, and clear handling in PluginRouterRegistryImpl.
- napcat-onebot types: Extend PluginRouterRegistry interface with no-auth API methods and document that authenticated APIs remain separate.
- napcat-webui-backend: Mount a new unauthenticated plugin route handler at /plugin/:pluginId/api that looks up the plugin router and dispatches requests to the plugin's no-auth router, returning appropriate errors when context or routes are missing.
- napcat-plugin-builtin: Add example no-auth endpoints (public/info and health) and update logger messages to reflect both auth and no-auth API paths.
- Bump napcat-types version to 0.0.16 and update napcat-plugin-builtin dependency accordingly.
These changes enable plugins to expose public endpoints (e.g. health checks or public metadata) under /plugin/{pluginId}/api/ while keeping existing authenticated APIs under /api/Plugin/ext/{pluginId}/.
Add a public route to serve plugin extension pages without auth and update related pieces accordingly. Backend: register GET /plugin/:pluginId/page/:pagePath to locate the plugin router, validate page and HTML file existence, and send the file (returns appropriate 4xx/5xx errors). Frontend: switch iframe and new-window URLs to the new unauthenticated route (remove webui_token usage). Builtin plugin: clarify page registration comment and add a log line for the extension page URL. Minor formatting whitespace tweaks in plugin manager type annotations.
Upgrade napcat-types to v0.0.15 and update the built-in plugin UI to test both filesystem and in-memory static resources. dashboard.html: clarify which plugin endpoints require auth, add buttons and a testMemoryResource() function that fetches an in-memory JSON resource, and add staticBase/memBase variables for non-auth static routes. napcat-webui-backend: return after 404 for missing memory files to stop further handling. (Lockfile updated accordingly.)
Introduce in-memory static file support and inter-plugin exports. Add MemoryStaticFile/MemoryFileGenerator types and expose staticOnMem in PluginRouterRegistry; router registry now tracks memory routes and exposes getters. Add getPluginExports to plugin manager adapters to allow plugins to call each other's exported modules. WebUI backend gains routes to serve /plugin/:pluginId/mem/* (memory files) and /plugin/:pluginId/files/* (plugin filesystem static) without auth. Update builtin plugin to demonstrate staticOnMem and inter-plugin call, and add frontend UI to open extension pages in a new window. Note: API router no longer mounts static filesystem routes — those are handled by webui-backend.
Add support for clearing CommonJS module cache when unloading plugins and reload plugins on install. PluginLoader now uses createRequire to access require.cache and exposes clearCache(pluginPath) to remove cached modules under the plugin directory; plugin manager calls this when unloading. Web UI backend install handlers now reload an existing plugin (with progress updates) instead of skipping registration, ensuring updated code/metadata take effect.
Backend: add CheckWebUIFontExist API and route; set --font-family-mono CSS variable in InitWebUi for aacute/custom/default. Improve webui font uploader: force saved filename to CustomFont, robustly clean old webui/CustomFont files, and log failures.
Frontend: add FileManager.checkWebUIFontExists; update theme settings to show upload UI only when 'custom' is selected, display uploaded status, attempt delete-before-upload, reload after actions, and adjust Accordion props. ColorPicker: enable pointer events on PopoverContent to allow dragging. applyFont now sets --font-family-mono for all modes.
Load local .env in napcat-shell and prioritize NAPCAT_WEBUI_SECRET_KEY across the app. Changes include:
- package.json: add build:shell:config script to build shell and copy env for dev builds.
- packages/napcat-develop/config/.env and loadNapCat.cjs: update default WEBUI secret to `napcatqq` and set the same env in the loader.
- packages/napcat-shell/napcat.ts: read config/.env into process.env at startup, rework pathWrapper/logger initialization, and minor formatting cleanups.
- packages/napcat-webui-backend/index.ts: InitWebUi now prefers NAPCAT_WEBUI_SECRET_KEY from the environment (updates config and logs the change); only generates a random token if no env override and current token is default.
These changes make it easier to override the WebUI token for development and ensure the token is propagated when building the shell for dev workflows.
* feat(webui): 新增配置全量备份与恢复功能。
* chore: Remove dependencies "archiver"
* feat(webui): 增加上传文件大小限制配置并优化上传处理
* Use memory-based zip import/export and multer
Replace disk-based zip handling with in-memory streaming to avoid temp files: remove unzipper/@types(unzipper) deps from package.json; update BackupConfig to stream-export configs with compressing.zip.Stream and to import by extracting uploaded zip buffer via compressing.zip.UncompressStream into in-memory Buffers. Backup of existing config is kept in-memory instead of copying to tmp, and imported files are written with path normalization checks. Router changed to use multer.memoryStorage() for uploads (remove dynamic tmp/disk upload logic and uploadSizeLimit usage). Also remove uploadSizeLimit from config schema.
* Revert "chore: Remove dependencies "archiver""
This reverts commit 733ec4c421dcb530d872a5ae1eb41bcf3bec78d5.
* Regenerate pnpm-lock.yaml (prune entries)
Regenerated pnpm-lock.yaml to reflect the current dependency resolution. This update prunes many removed/unused lock entries (notably archiver, unzipper and related @types, older/deprecated packages such as rimraf v2/fstream/bluebird, etc.) and removes platform 'libc' metadata from several platform-specific packages. There are no package.json changes; run `pnpm install` to sync your local node_modules with the updated lockfile.
---------
Co-authored-by: 手瓜一十雪 <nanaeonn@outlook.com>
Comment out the noisy '[NapCat Update] No pending updates found' log in UpdateNapCat.ts. Update frontend color choices: switch the plugin store action color from 'success' to 'default', and change the NewVersion chip and spinner from 'danger' to 'primary' in system_info.tsx. These tweaks reduce alarming red styling and quiet an unnecessary backend log.
Introduces backend API endpoints and frontend UI for managing SSL certificates, including viewing status, uploading, and deleting cert/key files. Adds a new SSL configuration tab in the dashboard, allowing users to enable HTTPS by providing PEM-formatted certificate and key, with changes taking effect after restart.
Introduces a plugin router registry for registering plugin-specific API routes, static resources, and extension pages. Updates the plugin manager and context to expose the router, and implements backend and frontend support for serving and displaying plugin extension pages in the WebUI. Also adds a demo extension page and static resource to the builtin plugin.
Implemented backend API and frontend UI for importing local plugin zip files. The backend now supports file uploads via a new /Plugin/Import endpoint using multer, and the frontend provides a button to upload and import plugins directly from the dashboard.
Prompt to register plugin manager if not loaded
Renames plugin_develop.ts to plugin-develop.ts for consistency. Updates the plugin import handler to prompt the user to register the plugin manager if it is not loaded, improving user experience and error handling.
Refactors the plugin manager by extracting configuration, loader, and type definitions into separate modules under the 'plugin' directory. Introduces a new PluginLoader class for scanning and loading plugins, and updates the main manager to use modularized logic and improved type safety. This change improves maintainability, separation of concerns, and extensibility for plugin management.
Introduces an 'imageDownloadProxy' config option to OneBot, allowing image downloads via a specified HTTP proxy. Updates the file download logic in napcat-common to use the undici library for proxy support, and propagates the new config through backend, frontend, and type definitions. Also adds undici as a dependency.
Introduces a reactive plugin configuration system with dynamic schema updates via server-sent events (SSE). Adds new fields and controller interfaces to the plugin manager, updates the built-in plugin to demonstrate dynamic config fields, and implements backend and frontend logic for real-time config UI updates. Also updates napcat-types to 0.0.10.
Introduces backend API and router for mirror management, including latency testing and custom mirror setting. Adds frontend components and controllers for mirror selection, speed testing, and integration into system info and plugin store pages, allowing users to select and test download/list mirrors interactively.
Enhanced setPluginStatus to support enabling/disabling plugins by both package name and dirname, improving robustness when plugins are not loaded. Also removed redundant directory name matching logic from findDirnameById in the web UI backend.
Register plugin after installation in PluginStore
Adds logic to immediately register a plugin with the plugin manager after installation, both in the standard and SSE install handlers. This ensures newly installed plugins are available without requiring a restart or manual reload.
Refactor plugin path handling in plugin manager
Simplifies plugin directory and data path resolution by using pluginPath from the plugin context instead of fileId. Streamlines plugin uninstall and reload logic, removing redundant file system scans and improving code clarity.
Refactor plugin API to use package id and improve UX
Standardized plugin management APIs and frontend to use 'id' (package name) instead of ambiguous 'name' or 'filename'. Added support for a 'plugin' display field in package.json and improved plugin store UI to show install/update status. Refactored backend and frontend logic for enabling, disabling, uninstalling, and configuring plugins to use consistent identifiers, and enhanced type definitions and documentation for better maintainability.
Introduces backend and frontend logic to manually register the plugin manager if not already loaded. Adds a new API endpoint and frontend UI prompt to guide users through registration after plugin installation when necessary.
Updated plugin manager and API to distinguish between plugin package name and directory name (dirname) for more robust plugin identification and path resolution. Adjusted context creation, status management, and API handlers to use package name for identification and dirname for filesystem operations. Also replaced console.error with console.log in builtin plugin for consistency.
Removed support for single-file plugins in OB11PluginMangerAdapter, simplifying plugin identification to use directory names as unique IDs. Updated related logic in the backend API to align with this change, ensuring consistent plugin management and status handling.
Removed support for single-file plugins in OB11PluginMangerAdapter, simplifying plugin identification to use directory names as unique IDs. Updated related logic in the backend API to align with this change, ensuring consistent plugin management and status handling.
Introduces a configuration UI schema and persistent config storage for the napcat-plugin-builtin. The plugin now loads and saves its configuration, supports dynamic prefix and reply toggling, and updates dependencies to napcat-types v0.0.6.
Introduces a unified plugin configuration schema and API in the backend, with endpoints for getting and setting plugin config. Updates the frontend to support plugin config modals, including a UI for editing plugin settings. Also adds support for uninstalling plugins with optional data cleanup, and updates dependencies to use napcat-types@0.0.5.
Introduces a new SSE-based plugin installation API for real-time progress updates and adds frontend support for selecting download mirrors, especially for GitHub-based plugins. Refactors backend plugin directory handling, improves logging, and updates the frontend to use the new API with user-selectable mirrors and progress feedback.
Replaces mock plugin store data with live fetching from a remote source using mirrors and caching. Implements actual plugin download, extraction, and installation logic in the backend. Updates frontend to call the new install API and display installation results, and removes unused icon/rating display from plugin cards.
Simplified PluginStoreItem types by removing unused fields and standardizing the minimum version property. Updated frontend plugin store page to use new categories ('官方', '工具', '娱乐', '其它') instead of download-based sorting, aligning with the revised type definitions.
* Add payload and return schemas to OneBot actions
Introduced explicit payloadSchema and returnSchema definitions for all OneBotAction classes using @sinclair/typebox. This improves type safety, API documentation, and validation for action payloads and return values. Also refactored method signatures and types for consistency across the codebase.
* Refactor payload schemas to use string IDs
Replaced Type.Union([Type.Number(), Type.String()]) with Type.String for group_id, user_id, and similar fields across all action payload schemas to standardize input types. Also made minor improvements to error handling, return types, and removed unused imports for better code clarity and consistency.
* Refactor type definitions and payload schemas in actions
Standardized type usage and improved type safety across multiple OneBot action files. Updated payload schemas to use string types for IDs and flags, refined return types, and enhanced message content typing. Added error handling for missing parameters in SetGroupTodo.
* Refactor type handling and improve message parsing
Updated several actions to use more precise type casting and type guards, improving type safety and clarity. Enhanced message parsing logic for forward messages and group/friend message history. Standardized return schemas and error handling for avatar and group portrait actions.
* Add napcat-schema package for OpenAPI generation
Introduces the napcat-schema package with scripts and configuration to auto-generate OpenAPI schemas for NapCat OneBot 11 actions. Refactors action handler export in napcat-onebot to support schema extraction.
* Add action examples and enhance action metadata
Introduced a centralized examples.ts file providing payload and return examples for all actions. Updated numerous action classes to include actionDescription, actionTags, payloadExample, and returnExample fields, improving API documentation and discoverability.
* Refactor action example imports and add example files
Moved action example data to dedicated 'examples.ts' files for each action category (extends, file, go-cqhttp, group, msg, system, user). Updated all action classes to import and use the new example modules, improving code organization and maintainability. Also added missing actionTags and actionDescription where appropriate.
* Update GetGroupMemberList.ts
* Add actionSummary and improve action metadata
Introduces the actionSummary property to OneBotAction and updates all action classes to provide concise summaries and improved descriptions. Refactors example imports for better modularity, adds new example files for guild and packet actions, and updates the OpenAPI schema generator to use the new summary and improved descriptions. This enhances API documentation clarity and consistency.
* Enhance action metadata and add examples for new actions
Added actionSummary, actionDescription, and actionTags to multiple OneBot actions for improved API documentation. Introduced payload and response examples for new actions (GetDoubtFriendsAddRequest, SetDoubtFriendsAddRequest) in a new examples.ts file. Also removed unused imports from several files for code clarity.
* Refactor action examples and enhance metadata
Replaced generic ActionExamples imports with more specific examples modules (FileActionsExamples, GroupActionsExamples, GoCQHTTPActionsExamples) across file, group, and go-cqhttp actions. Added or updated actionSummary, actionDescription, actionTags, payloadExample, and returnExample properties for improved API documentation and clarity.
* Refactor extends actions to use new examples module
Replaced imports of ActionExamples with ExtendsActionsExamples in all extends actions. Updated action summary, description, tags, and example references for consistency and clarity across actions. This improves maintainability and aligns with the new examples structure.
* Add action metadata to OneBot action classes
Added or updated actionSummary, actionTags, payloadExample, and returnExample properties for all OneBot action classes in the napcat-onebot package. This improves API documentation and discoverability by providing concise summaries, categorization tags, and usage examples for each action.
* Refactor OpenAPI schema generation to 3.0.1 format
Updated the OpenAPI schema output to use version 3.0.1, restructured tags, responses, and examples for better clarity and compatibility, and simplified output file locations. Also removed unused scripts from package.json.
* Fix SendPokePayloadSchema type definitions
Corrected the type definitions for user_id and target_id to only allow strings, and fixed a syntax error in group_id. This ensures payload validation is consistent and accurate.
Refactor fileset ID API response and schema handling
Updated GetFilesetId action to return a structured object with fileset_id and adjusted its return schema accordingly. Improved frontend TypeBox schema parsing to support allOf (intersection) merging and updated API debug component to construct response schemas in a more robust way for object recognition.
Refactor OneBot API schema handling to use TypeBox
Replaces Zod-based static API schema definitions with dynamic fetching of schemas from the backend using TypeBox. Removes legacy static schema files, updates frontend API debug components to use TypeBox utilities, and adds @sinclair/typebox as a dependency. Backend now exposes a /schemas endpoint for all OneBot actions. Various schema and description fields are updated for clarity and consistency.
* Remove OneBot API navigation list component
Deleted nav_list.tsx from the onebot/api components, removing the OneBotApiNavList React component and its related logic. This may be part of a refactor or cleanup to eliminate unused or redundant UI code.
* Add action tags to OneBot API schema and update tag name
Included the 'tags' property in the OneBot API schema for both backend and frontend, allowing actions to be categorized. Also updated the action tag from '群扩展' to '群组扩展' in SetGroupSign for consistency.
* Add napcat-types package for unified type exports
Introduced the napcat-types package to aggregate and re-export all types, enums, and classes from napcat-core and napcat-onebot. Added external module shims, test files, and configuration for type-only distribution. Updated core and onebot packages to improve export granularity and fixed import paths for better modularity.
* Move external-shims.d.ts to files in tsconfig
external-shims.d.ts was moved from the include array to the files array in tsconfig.json to ensure it is always included explicitly. This change clarifies the intent and may help with TypeScript's file resolution.
Refactor napcat-types package and update plugin deps
Refactored napcat-types to provide more accurate shims, added real type dependencies, and improved build/test scripts. Updated napcat-plugin and napcat-plugin-builtin to depend on napcat-types instead of napcat-onebot. Adjusted imports in affected packages to use napcat-types, and updated pnpm-lock.yaml accordingly.
Add build and test scripts to napcat-types package
Introduced 'build' and 'test' scripts in the napcat-types package.json for easier development and testing. Also updated dependencies in the lockfile.
* 完善部分api描述
* Remove unused statusText constant
Deleted the unused statusText constant from FetchCustomFace.ts to clean up the code.
* Bump napcat-types version to 0.0.2
Updated the package version in package.public.json from 0.0.1 to 0.0.2.
Update napcat-types package metadata and dependencies
Set package as public by changing 'private' to false. Move 'napcat-core' and 'napcat-onebot' from dependencies to devDependencies, and remove 'compressing' from dependencies.
Add public packaging and build script for napcat-types
Introduces package.public.json and a copy-dist.mjs script to automate copying metadata and README into the dist folder for publishing. Updates build script in package.json to use the new copy step. Adds initial package.json and README for napcat-types.
Update scripts in napcat-types package configs
Added a publish script to package.json and removed scripts from package.public.json to streamline configuration and avoid duplication.
* Update publish script to use npm in dist directory
Changed the publish script to run 'npm publish' from the 'dist' directory instead of using 'pnpm publish --filter napcat-types'. This ensures the published package uses the built output.
* Update pnpm-lock.yaml dependencies
Removed 'compressing' from dependencies and cleaned up libc fields for various platform-specific packages. This streamlines the lock file and may improve cross-platform compatibility.
* Add workflow to publish OpenAPI schema to NapCatDocs
Introduces a new 'publish-schema' job in the auto-release workflow. This job builds the napcat-schema package, copies the generated OpenAPI schema to the NapCatDocs repository under a versioned path, and commits the update. Automates schema publishing on release events.
* AI修正部分api文档
* Update OpenAPI version and use dynamic version from napcat-common
Changed OpenAPI spec version to 3.1.0 and replaced the hardcoded API version with napCatVersion from napcat-common. Added napcat-common as a dependency in package.json.
* Update napcat-schema build and OpenAPI version
Renamed the build script from build:schema to build:openapi in napcat-schema and updated the workflow to use the new script. Changed OpenAPI version from 3.1.0 to 3.0.1 in the schema generator. Added napcat-vite as a dependency and integrated its version plugin into the Vite config.
* 暂时OK
* Refactor action examples structure and imports
Moved action example files into a new 'example' directory and updated all imports accordingly. Removed the monolithic 'examples.ts' and redefined ActionExamples in OneBotAction.ts to only include common error codes. This improves code organization and maintainability.
* Fix type for rate limiter middleware in router
Casts the rate limiter middleware to RequestHandler to resolve type compatibility issues with Express router middleware.
* Add OB11 message segment schemas and update SendMsg
Introduces a comprehensive message segment schema (OB11) in a new file, refactors SendMsg payload to use the new OB11MessageMixTypeSchema, and updates related type definitions for improved type safety and extensibility in message handling.
* Refactor OB11 message types to use TypeBox schemas
Migrates all OB11 message segment and message type definitions from interface/enums to TypeBox schemas in types/message.ts. Removes the now-redundant message-segment-schema.ts file and updates imports to use the new schema-based types. This unifies type validation and TypeScript types, improving maintainability and consistency.
---------
Co-authored-by: pohgxz <Ttjonne@outlook.com>
Replaces the 'disableNonLANAccess' option with a more flexible access control system supporting 'none', 'whitelist', and 'blacklist' modes, along with IP list and X-Forwarded-For support. Updates backend API, config schema, middleware, and frontend UI to allow configuration of access control mode, IP whitelist/blacklist, and X-Forwarded-For handling. Removes legacy LAN-only access logic and updates types accordingly.
Implemented plugin store API endpoints and types in the backend, including mock data and handlers for listing, detail, and install actions. Added plugin store page, card component, and related logic to the frontend, with navigation and categorized browsing. Updated plugin manager controller and site config to support the new plugin store functionality.
Standardized type annotations for interfaces in user.ts and improved type safety in webapi.ts. Updated all Express router initializations to explicitly declare the Router type. Added missing RequestHandler typings in uploader modules for better type checking.
