- Limit popup width to a configurable range (40‑80 characters) and ensure a minimum width of 1.
- Preserve original width when the terminal is narrower than the minimum.
- Clamp the y coordinate to the top of the area to avoid negative positioning.
- Introduce `system_status_message` helper to determine the message shown in the system/status pane.
- Calculate wrapped line count based on available width, clamp visible rows to 1–5, and set the layout constraint dynamically.
- Update `render_system_output` to accept the pre‑computed message, choose color based on error prefix, and render each line individually, defaulting to “Ready” when empty.
- Adjust UI layout to use the new dynamic constraint for the system/status section.
- Map git badges and cleanliness states to specific `Color` values and modifiers.
- Apply these colors to file icons, filenames, and markers in the UI.
- Propagate the most relevant dirty badge from child nodes up to parent directories.
- Extend the help overlay with a “GIT COLORS” section describing the new color legend.
- Update help UI to show “Ctrl+↑/↓ → resize chat/thinking split”.
- Introduce `ensure_ratio_bounds` and `nudge_ratio` on `LayoutNode` to clamp and adjust split ratios.
- Ensure vertical split favors the thinking panel when it becomes focused.
- Add `adjust_vertical_split` method in `ChatApp` and handle Ctrl+↑/↓ in normal mode to modify the split and update status messages.
- Update help UI to show “Ctrl+←/→ → resize files panel”.
- Change `set_file_panel_width` to return the clamped width.
- Implement Ctrl+←/→ handling in keyboard input to adjust the files panel width, update status messages, and respect panel collapse state.
- Introduced a new “HELP & QUICK COMMANDS” section with bold header and shortcuts for toggling the help overlay and opening the files panel.
- Updated command help text to “Open the help overlay”.
- Extended onboarding and tutorial status lines to display the help shortcut.
- Modified help command handling to set the status to “Help” and clear errors.
- Introduce “Ctrl+C twice” shortcut for quitting the application and display corresponding help line.
- Rename and clarify session‑related commands (`:session save`) and add short aliases (`:w[!]`, `:q[!]`, `:wq[!]`) with updated help entries.
- Adjust quit help text to remove `:q, :quit` redundancy and replace with the new quick‑exit hint.
- Update UI key hint to show only “Esc” for cancel actions.
- Implement double‑Ctrl+C detection in `ChatApp` using `DOUBLE_CTRL_C_WINDOW`, track `last_ctrl_c`, reset on other keys, and show status messages prompting the second press.
- Minor wording tweaks in help dialogs and README to reflect the new command syntax and quick‑exit behavior.
- Updated command specs: added `w`, `write`, `wq`, `x`, and `session save` with proper descriptions.
- Introduced `SaveStatus` enum and helper methods for path display and buffer labeling.
- Implemented `update_paths` in `Workspace` to keep title in sync with file paths.
- Added comprehensive `save_active_code_buffer` and enhanced `close_active_code_buffer` logic, including force‑close via `!`.
- Parsed force flag from commands (e.g., `:q!`) and routed commands to new save/close workflows.
- Integrated session save subcommand with optional description generation.
- Adjust layout constraint from 5 to 4 lines to match 2 lines of content plus borders.
- Refactor file focus key handling to toggle the file panel state and set status messages (“Files panel shown” / “Files panel hidden”) instead of always expanding and using a static status.
- Render the app title with version as the block title instead of a separate header widget.
- Compute `content_area` via `main_block.inner` and use it for file panel, main area, model info panel, and toast rendering.
- Remove header constraints and the `render_header` function, simplifying the layout.
- Add early exit when `content_area` has zero width or height to avoid rendering errors.
Add `APP_VERSION` constant derived from `CARGO_PKG_VERSION` and update the header rendering to display the version (e.g., “🦉 OWLEN v1.2.3 – AI Assistant”).
Introduce `McpCommand` enum and handlers in `owlen-cli` to manage MCP server registrations, including adding, listing, and removing servers across configuration scopes. Add scoped configuration support (`ScopedMcpServer`, `McpConfigScope`) and OAuth token handling in core config, alongside runtime refresh of MCP servers. Implement toast notifications in the TUI (`render_toasts`, `Toast`, `ToastLevel`) and integrate async handling for session events. Update config loading, validation, and schema versioning to accommodate new MCP scopes and resources. Add `httpmock` as a dev dependency for testing.
- Introduce `IconMode` in core config (default Auto) and bump schema version to 1.4.0.
- Add `FileIconSet`, `IconDetection`, and `FileIconResolver` to resolve per‑file icons with configurable fallbacks and environment variable `OWLEN_TUI_ICONS`.
- Export resolver types from `owlen-tui::state::file_icons`.
- Extend `ChatApp` with `file_icons` field, initialize it from config, and expose via `file_icons()` accessor.
- Append system status line showing selected icon set and detection source.
- Implement breadcrumb construction (`repo > path > file`) and display in code pane headers.
- Render icons in file tree, handle unsaved file markers, hidden files, and Git decorations with proper styling.
- Add helper `collect_unsaved_relative_paths` and tree line computation for visual guides.
- Provide `Workspace::panes()` iterator for unsaved tracking.
- Update UI imports and tests to cover new breadcrumb feature.
- Export `PaletteGroup` and `PaletteSuggestion` to represent suggestion metadata.
- Implement command history with deduplication, capacity limit, and recent‑command suggestions.
- Enhance dynamic suggestion logic to include history, commands, models, and providers with fuzzy ranking.
- Add UI rendering for grouped suggestions, header with command palette label, and footer instructions.
- Update help text with new shortcuts (Ctrl+P, layout save/load) and expose new agent/layout commands.
Add `focus_beacon_span`, `panel_title_spans`, `panel_hint_style`, and `panel_border_style` utilities to centralize panel header, hint, border, and beacon rendering. Integrate these helpers across all UI panels (files, chat, thinking, agent actions, input, status bar) and update help text. Extend `Theme` with new color fields for beacons, pane headers, and hint text, providing defaults for all built‑in themes. Include comprehensive unit tests for the new styling functions.
Introduce `FileTreeState` for managing a navigable file hierarchy with Git decorations, filtering, and cursor/scroll handling.
Add `RepoSearchState` and related types to support asynchronous ripgrep‑backed repository searches, including result aggregation, pagination, and UI interaction.
- Introduce `RoleLabelDisplay` enum (inline, above, none) and integrate it into UI rendering and message formatting.
- Replace `show_role_labels` boolean with `role_label_mode` across config, formatter, session, and TUI components.
- Add `syntax_highlighting` boolean to UI settings with default `false` and support in message rendering.
- Update configuration schema version to 1.3.0 and provide deserialization handling for legacy boolean values.
- Extend theme definitions with code block styling fields (background, border, text, keyword, string, comment) and default values in `Theme`.
- Adjust related modules (`formatting.rs`, `ui.rs`, `session.rs`, `chat_app.rs`) to use the new settings and theme fields.
- Added `show_cursor_outside_insert` (default false) to `UiSettings` and synced it from config.
- Cursor rendering now follows `cursor_should_be_visible`, allowing visibility outside insert mode based on the new setting.
- Replaced `textwrap::wrap` with `wrap_unicode`, which uses Unicode break properties for proper CJK and emoji handling.
- Added `grayscale-high-contrast.toml` theme, registered it in theme loading, and updated README and tests.
- Add placeholder span showing shortcuts (i:Insert, m:Model, ?:Help, : Command) in the UI footer.
- Insert help section describing Enter key behavior in normal and insert modes.
- Extend F1 help shortcut to also trigger on “?” key (with no or Shift modifier).
- Introduce `Theme` import and pass a cloned `theme` instance to UI helpers (e.g., `render_editable_textarea`).
- Remove direct `Color` usage; UI now derives colors from `Theme` fields for placeholders, selections, ReAct components (thought, action, input, observation, final answer), status badges, operating mode badges, and model info panel.
- Extend `Theme` with new color fields for agent ReAct stages, badge foreground/background, and operating mode colors.
- Update rendering logic to apply these theme colors throughout the TUI (input panel, help text, status lines, model selection UI, etc.).
- Adjust imports to drop unused `Color` references.
- Introduce `MessageRenderContext` and `MessageCacheEntry` for caching wrapped lines per message.
- Implement `render_message_lines_cached` using cache, invalidating on updates.
- Add role/style helpers and content hashing for cache validation.
- Throttle UI redraws in the main loop during active streaming (50 ms interval) and adjust idle tick timing.
- Update drawing logic to use cached rendering and manage draw intervals.
- Remove unused `role_color` function and adjust imports accordingly.
- Introduce `ModelPaletteEntry` and re‑export it for external use.
- Extend `CommandPalette` with dynamic sources (models, providers) and methods to refresh suggestions based on `:model` and `:provider` prefixes.
- Implement fuzzy matching via `match_score` and subsequence checks for richer suggestion ranking.
- Add `provider` command spec and completions.
- Update UI to display “Model (Provider)” in the header and use the new active model label helper.
- Wire catalog updates throughout `ChatApp` (model palette entries, command palette refresh on state changes, model picker integration).
Introduce `ui.scrollback_lines` (default 2000) to cap the number of chat lines kept in memory, with `0` disabling trimming. Implement automatic trimming of older lines, maintain a scroll offset, and show a “↓ New messages (press G)” badge when new messages arrive off‑screen. Update core UI settings, TUI rendering, chat app state, migrations, documentation, and changelog to reflect the new feature.
- Introduce `ui.input_max_rows` (default 5) to control how many rows the input panel expands before scrolling.
- Bump `CONFIG_SCHEMA_VERSION` to **1.2.0** and update migration documentation.
- Update `configuration.md` and migration guide to describe the new setting.
- Adjust TUI height calculation to respect `input_max_rows` and add horizontal scrolling support for long lines.
- Add `unicode-segmentation` dependency for proper grapheme handling.
Introduce `DetailedModelInfo` and `ModelInfoRetrievalError` structs for richer model data.
Add `ModelDetailsCache` with TTL‑based storage and async API for get/insert/invalidate.
Extend `OllamaProvider` to fetch, cache, refresh, and list detailed model info.
Expose model‑detail methods in `Session` for on‑demand and bulk retrieval.
Add `ModelInfoPanel` widget to display detailed info with scrolling support.
Update TUI rendering to show the panel, compute viewport height, and render model selector labels with parameters, size, and context length.
Adjust imports and module re‑exports accordingly.
- Switch to `ollama-rs` crate for chat, model listing, and streaming.
- Remove custom request building, authentication handling, and debug logging.
- Drop unsupported tool conversion; now ignore tool descriptors with a warning.
- Refactor model fetching to use local model info and optional cloud details.
- Consolidate error mapping via `map_ollama_error`.
- Update health check to use the new HTTP client.
- Delete obsolete `provider_interface.rs` test as the provider interface has changed.
Deletes the `owlen-ollama` Cargo.toml and source files, fully removing the Ollama provider from the workspace. This aligns the project with the MCP‑only architecture and eliminates direct provider dependencies.
- Introduce `show_onboarding` UI setting (default true) and persist its state after first launch.
- Show onboarding status line and system status on initial run; fallback to normal status thereafter.
- Implement `show_tutorial` method displaying keybinding tips and system status.
- Register `:tutorial` command in command palette.
- Add migration documentation explaining `schema_version` update and deprecation of `agent.max_tool_calls`.
- Update README with description of the new tutorial command.
- Detect terminal color support and automatically switch to the new `ansi_basic` theme when only 16‑color support is available.
- Introduce `OfflineProvider` that supplies a placeholder model and friendly messages when no providers are reachable, keeping the TUI usable.
- Add `CONFIG_SCHEMA_VERSION` (`1.1.0`) with schema migration logic and default handling in `Config`.
- Update configuration saving to persist the schema version and ensure defaults.
- Register the `ansi_basic` theme in `theme.rs`.
- Extend `ChatApp` with `set_status_message` to display custom status lines.
- Update documentation (architecture, Vim mode state machine) to reflect new behavior.
- Add async‑trait and futures dependencies required for the offline provider implementation.
- Export `LLMProvider` from `owlen-core` and replace public `Provider` re-exports.
- Convert `OllamaProvider` to implement the new `LLMProvider` trait with associated future types.
- Adjust imports and trait bounds in `remote_client.rs` to use the updated types.
- Add comprehensive provider interface tests (`provider_interface.rs`) verifying router routing and provider registry model listing with `MockProvider`.
- Align dependency versions across workspace crates by switching to workspace-managed versions.
- Extend CI (`.woodpecker.yml`) with a dedicated test step and generate coverage reports.
- Update architecture documentation to reflect the new provider abstraction.
This commit fixes 12 categories of errors across the codebase:
- Fix owlen-mcp-llm-server build target conflict by renaming lib.rs to main.rs
- Resolve ambiguous glob re-exports in owlen-core by using explicit exports
- Add Default derive to MockMcpClient and MockProvider test utilities
- Remove unused imports from owlen-core test files
- Fix needless borrows in test file arguments
- Improve Config initialization style in mode_tool_filter tests
- Make AgentExecutor::parse_response public for testing
- Remove non-existent max_tool_calls field from AgentConfig usage
- Fix AgentExecutor::new calls to use correct 3-argument signature
- Fix AgentResult field access in agent tests
- Use Debug formatting instead of Display for AgentResult
- Remove unnecessary default() calls on unit structs
All changes ensure the project compiles cleanly with:
- cargo check --all-targets ✓
- cargo clippy --all-targets -- -D warnings ✓
- cargo test --no-run ✓
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
This commit completes the Phase 10 migration to MCP-only architecture by
removing all legacy mode code paths and configuration options.
**Breaking Changes:**
- Removed `McpMode` enum from configuration system
- Removed `mode` setting from `[mcp]` config section
- MCP architecture is now always enabled (no option to disable)
**Code Changes:**
- Simplified `McpSettings` struct (now a placeholder for future options)
- Updated `McpClientFactory` to remove legacy mode branching
- Always use MCP architecture with automatic fallback to local client
- Added test infrastructure: `MockProvider` and `MockMcpClient` in test_utils
**Documentation:**
- Created comprehensive v0.x → v1.0 migration guide
- Added CHANGELOG_v1.0.md with detailed technical changes
- Documented common issues (cloud model 404s, timeouts, API key setup)
- Included rollback procedures and troubleshooting steps
**Testing:**
- All 29 tests passing
- Fixed agent tests to use new mock implementations
- Updated factory test to reflect new behavior
This completes the 10-phase migration plan documented in .agents/new_phases.md,
establishing Owlen as a production-ready MCP-only TUI application.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Phase 10 "Cleanup & Production Polish" is now complete. All LLM
interactions now go through the Model Context Protocol (MCP), removing
direct provider dependencies from CLI/TUI.
## Major Changes
### MCP Architecture
- All providers (local and cloud Ollama) now use RemoteMcpClient
- Removed owlen-ollama dependency from owlen-tui
- MCP LLM server accepts OLLAMA_URL environment variable for cloud providers
- Proper notification handling for streaming responses
- Fixed response deserialization (McpToolResponse unwrapping)
### Code Cleanup
- Removed direct OllamaProvider instantiation from TUI
- Updated collect_models_from_all_providers() to use MCP for all providers
- Updated switch_provider() to use MCP with environment configuration
- Removed unused general config variable
### Documentation
- Added comprehensive MCP Architecture section to docs/architecture.md
- Documented MCP communication flow and cloud provider support
- Updated crate breakdown to reflect MCP servers
### Security & Performance
- Path traversal protection verified for all resource operations
- Process isolation via separate MCP server processes
- Tool permissions controlled via consent manager
- Clean release build of entire workspace verified
## Benefits of MCP Architecture
1. **Separation of Concerns**: TUI/CLI never directly instantiates providers
2. **Process Isolation**: LLM interactions run in separate processes
3. **Extensibility**: New providers can be added as MCP servers
4. **Multi-Transport**: Supports STDIO, HTTP, and WebSocket
5. **Tool Integration**: MCP servers expose tools to LLMs
This completes Phase 10 and establishes a clean, production-ready architecture
for future development.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
This commit completes Phase 10 of the MCP migration by removing all
direct provider usage from CLI/TUI and enforcing MCP-first architecture.
## Changes
### Core Architecture
- **main.rs**: Replaced OllamaProvider with RemoteMcpClient
- Uses MCP server configuration from config.toml if available
- Falls back to auto-discovery of MCP LLM server binary
- **agent_main.rs**: Unified provider and MCP client to single RemoteMcpClient
- Simplifies initialization with Arc::clone pattern
- All LLM communication now goes through MCP protocol
### Dependencies
- **Cargo.toml**: Removed owlen-ollama dependency from owlen-cli
- CLI no longer knows about Ollama implementation details
- Clean separation: only MCP servers use provider crates internally
### Tests
- **agent_tests.rs**: Updated all tests to use RemoteMcpClient
- Replaced OllamaProvider::new() with RemoteMcpClient::new()
- Updated test documentation to reflect MCP requirements
- All tests compile and run successfully
### Examples
- **Removed**: custom_provider.rs, basic_chat.rs (deprecated)
- **Added**: mcp_chat.rs - demonstrates recommended MCP-based usage
- Shows how to use RemoteMcpClient for LLM interactions
- Includes model listing and chat request examples
### Cleanup
- Removed outdated TODO about MCP integration (now complete)
- Updated comments to reflect current MCP architecture
## Architecture
```
CLI/TUI → RemoteMcpClient (impl Provider)
↓ MCP Protocol (STDIO/HTTP/WS)
MCP LLM Server → OllamaProvider → Ollama
```
## Benefits
- ✅ Clean separation of concerns
- ✅ CLI is protocol-agnostic (only knows MCP)
- ✅ Easier to add new LLM backends (just implement MCP server)
- ✅ All tests passing
- ✅ Full workspace builds successfully
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
Remove the separate owlen-code binary as code assistance functionality
is now integrated into the main application through the mode consolidation system.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Fix ACTION_INPUT regex to properly capture multiline JSON responses
- Changed from stopping at first newline to capturing all remaining text
- Resolves parsing errors when LLM generates formatted JSON with line breaks
- Enhance tool schemas with detailed descriptions and parameter specifications
- Add comprehensive Message schema for generate_text tool
- Clarify distinction between resources/get (file read) and resources/list (directory listing)
- Include clear usage guidance in tool descriptions
- Set default model to llama3.2:latest instead of invalid "ollama"
- Add parse error debugging to help troubleshoot LLM response issues
The agent infrastructure now correctly handles multiline tool arguments and
provides better guidance to LLMs through improved tool schemas. Remaining
errors are due to LLM quality (model making poor tool choices or generating
malformed responses), not infrastructure bugs.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-Authored-By: Claude <noreply@anthropic.com>
- Introduce `owlen-mcp-llm-server` crate with RPC handling, `generate_text` tool, model listing, and streaming notifications.
- Add `RpcNotification` struct and `MODELS_LIST` method to the MCP protocol.
- Update `owlen-core` to depend on `tokio-stream`.
- Adjust Ollama provider to omit empty `tools` field for compatibility.
- Enhance `RemoteMcpClient` to locate the renamed server binary, handle resource tools locally, and implement the `Provider` trait (model listing, chat, streaming, health check).
- Add new crate to workspace `Cargo.toml`.
- Added a `tool_output` color to the `Theme` struct.
- Updated all built-in themes to include the new color.
- Modified the TUI to use the `tool_output` color for rendering tool output.
- Added a `tool_output` color to the `Theme` struct.
- Updated all built-in themes to include the new color.
- Modified the TUI to use the `tool_output` color for rendering tool output.
- Added a `tool_output` color to the `Theme` struct.
- Updated all built-in themes to include the new color.
- Modified the TUI to use the `tool_output` color for rendering tool output.
