Refactor subagent model configuration from nested modelConfig object to a simple model string field for better UX and clarity.
Changes:
- Replace modelConfig object with model string in SubagentConfig interface
- Add model-selection.ts utility for parsing and validating model selectors
- Support 'inherit' keyword and bare model IDs (e.g., 'glm-5', 'claude-sonnet-4-6')
- Maintain backward compatibility by parsing legacy modelConfig frontmatter
- Update validation to reject cross-provider authType-prefixed selectors
- Update SDK types (TypeScript and Java) to reflect new schema
- Add comprehensive tests for model selection and validation
- Update documentation with model selection examples
Breaking changes:
- modelConfig.frontmatter field deprecated in favor of model field
- Cross-provider model selectors (e.g., 'openai:gpt-4') not supported for subagents
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
Key changes:
1. Remove LspLanguageDetector — LSP is now fully config-driven via
.lsp.json or extensions. No more auto-detected built-in presets
that fail when the server binary is missing.
2. Add ensureDocumentOpen — send textDocument/didOpen before every
document-level LSP request (definitions, references, hover,
documentSymbol, implementations, prepareCallHierarchy, diagnostics,
codeActions). This fixes the root cause of most methods returning
empty results (Issue #2106, #1873).
3. Add retry mechanism for slow servers — when a freshly opened
document yields empty results on non-TypeScript servers (jdtls,
clangd, pylsp), wait 2s and retry once. This mirrors the existing
retry logic in workspaceSymbols.
4. Handle LSP 3.17 WorkspaceSymbol format — location.range is now
optional in normalizeSymbolResult, fixing jdtls workspace symbol
responses that omit the range field.
5. Improve workspace symbol warmup — for non-TypeScript servers, open
a workspace file before the first workspace/symbol request and
retry on empty results after a warmup delay.
6. Track warmup-opened URIs — warmupTypescriptServer now returns the
opened URI, which is registered with ensureDocumentOpen to prevent
duplicate didOpen notifications.
Closes#2106, closes#1873
Made-with: Cursor
Add comprehensive documentation for the Agent Arena feature, covering
usage, configuration, best practices, troubleshooting, and limitations.
Update navigation metadata to include the new page.
This enables users to discover and learn about the multi-model comparison
capability for competitive task execution.
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
Move the Docker sandbox Java documentation from docs-site to the
appropriate location in docs/users/features/sandbox.md under the
Troubleshooting section.
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
- Increase description warning threshold from 500 to 1,000 characters
- Change system prompt 10,000 char limit from error to warning
- Remove intermediate 5,000 char warning threshold for system prompts
- Update documentation to reflect soft warning behavior
This provides more flexibility for users while still guiding them
toward better practices.
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
- Add /context command to the Interface and Workspace Control table
- Document the context window usage breakdown feature
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
- Change description field limit from 301 to 300 characters
- Verified limits from source code in CreationSummary.tsx
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
- Corrected the file path in commands.md mapping table from
'<project>/commands/git/commit.md' to '<project>/.qwen/commands/git/commit.md'
- Removed trailing blank line in sandbox.md code block
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
- Document the 301 character limit for description field
- Document the 10,000 character limit for system prompt
Co-authored-by: Qwen-Coder <qwen-coder@alibabacloud.com>
- Delete listCommand and refreshCommand from mcpCommand.ts
- Update subCommands to only include manageCommand and authCommand
- Update documentation to reference MCP management dialog instead of CLI commands
- Simplify mcp command description to focus on management dialog and OAuth auth
Note: i18n strings for deprecated commands are kept for backward compatibility
- remove configuration options such as lsp.enabled, lsp.allowed, lsp.excluded, etc. from settings.json schema
- Delete lspSettingsSchema.ts files and associated JSON schema definitions
- Removed VS Code settings loading function, no longer merge. vscode/settings.json configuration
- Updated LSP documentation to reflect new configurations and experimental flags
-remove allow/exclude parameters in NativeLspService constructor
- Create new LspConfigLoader classes to handle LSP configuration loading and merging
- Updated debug guide to match the new configuration mechanism
- Simplify loadCliConfig functions, remove startLsp options
- Reconstruct the configuration loading process to remove duplicate configuration merge logic
- Add LspConfigLoader classes to implement configuration parsing and merging functions
- Add tools.experimentalSkills setting in settingsSchema
- Read default from settings in config. ts
- Add --skills as shorter alias for --experimental-skills
- Update documentation for new setting
Fixes#1493
Fixes#1359.
Default container sandboxing on Linux to use host UID/GID so qwen runs under a user that matches the mounted home directory and persists auth/settings in ~/.qwen.
Also gate the informational log behind DEBUG/DEBUG_MODE and clarify docs about Linux UID/GID mapping and ~/.qwen persistence.
