Add flex flex-col min-h-0 to the skills container so overflow-y-auto
in SkillsManager can properly trigger scrolling.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Providers API now detects ANTHROPIC_API_KEY, ANTHROPIC_AUTH_TOKEN, and
ANTHROPIC_BASE_URL from environment. ProviderManager shows a banner
when env vars are detected, informing users they're used automatically
when no provider is active.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Slash commands:
- Fix skills not appearing (removed non-existent `enabled` filter)
- Add badge/capsule UI for prompt-type commands (/compact, /review, etc.)
- SDK-native commands sent directly; CLI-only commands expanded to prompts
- /clear now resets SDK session and clears database messages
- /cost shows actual aggregated token usage statistics
- Keyboard: Backspace/Escape to remove badge, arrow keys auto-scroll popover
- Add /memory command for editing CLAUDE.md
Multi-provider management:
- CRUD API for API providers with SQLite storage
- ProviderManager UI with quick presets (GLM, Kimi, MiniMax, etc.)
- Dynamic model labels based on active provider
- Environment variable injection with conflict resolution
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The collectStreamResponse() function only collected text events,
discarding tool_use and tool_result events. When a conversation was
resumed, all tool call context was lost — the user only saw the
text portions of Claude's responses.
Changes:
- Collect tool_use and tool_result SSE events as MessageContentBlock
entries alongside text blocks
- Flush accumulated text into a text block before each tool_use to
preserve the interleaved ordering (text → tool_use → tool_result →
text → ...)
- Store messages with tool calls as JSON arrays of MessageContentBlock,
using the existing type definitions in types/index.ts
- Text-only messages remain as plain strings for backward compatibility
with existing message rendering (parseMessageContent handles both)
- Add 10 unit tests covering parsing, serialization, round-tripping,
edge cases, and backward compatibility (all passing)
This enables future improvements like:
- Rendering tool call history in resumed conversations
- Showing which files were read/written in previous turns
- Exporting full conversation logs with tool context
- Fix Windows Claude CLI not found when installed as .exe or without
extension in .local\bin (closes#6). Now searches all extensions
(.cmd, .exe, .bat, and bare) for every candidate directory.
- Fix Extensions/Skills list not scrollable when many items exist.
Change content container to overflow-hidden so internal ScrollArea
works correctly. Add overflow-auto to McpManager for its own scroll.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
- Add src/lib/platform.ts: cross-platform Claude CLI discovery
(candidate paths, .cmd/.bat shell handling, where/which fallback)
- Refactor claude-status route and claude-client to use shared platform utils
- Electron main: Windows process spawning (windowsHide), PATH construction
with AppData/npm dirs, titleBarOverlay, .ico icon support
- electron-builder.yml: add win/nsis build targets
- package.json: add electron:pack:win script
- FolderPicker: drive switcher dropdown for navigating between drives
- SkillsManager: fix duplicate key warning with filePath fallback
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
When navigating to a drive root on Windows (e.g. C:\), path.dirname
returns itself, disabling the "go up" button and trapping the user on
that drive. This adds a compact dropdown next to the up button that
lists all available drives (auto-detected via fs.accessSync), allowing
quick switching between drives. macOS/Linux are unaffected (empty array,
dropdown not rendered).
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
1. Skills API: accept optional cwd query param so project-level skills
are scanned from the session's working directory, not the standalone
server's process.cwd().
2. SDK env: explicitly set HOME so Claude Code can resolve ~/.claude/
for global skills discovery in packaged builds.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Add a new "API Configuration" section in settings for users who use
custom Anthropic-compatible APIs (e.g. via ai-claude-start). Users
can optionally set ANTHROPIC_AUTH_TOKEN and ANTHROPIC_BASE_URL in
CodePilot's settings UI. These are stored in SQLite and injected
into the Claude Code subprocess env only when configured.
Existing authentication methods (claude login, shell env vars) are
not affected - the API config is purely additive.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Return full error stack traces from the sessions API and display
the actual server error in the frontend, instead of a generic
"Failed to create session" message.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
The claude-status endpoint now searches common installation paths
(/usr/local/bin, /opt/homebrew/bin, ~/.npm-global/bin, ~/.local/bin,
~/.claude/bin) and uses `which` with expanded PATH as fallback,
fixing "Disconnected" status in packaged builds.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
Full-featured desktop application built with Next.js + Electron:
- Streaming chat with Claude Code via Agent SDK
- Session management with SQLite persistence
- Project file tree and file preview
- MCP server and custom skills management
- Permission controls with Code/Plan/Ask modes
- Dark/light theme, token tracking
- macOS DMG packaging with electron-builder
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
