Business copilot in the menu bar - Electron desktop application for time tracking, task management, activity monitoring, and AI-powered assistance
This is an Electron desktop app, not an MCP server. The codebase contains tool definitions in src/renderer/lib/chat/toolConfigs.ts but no MCP protocol implementation is visible. Tools are defined for a local chat interface, not exposed via MCP transport. No input schemas, parameter descriptions, or output schemas are visible in the provided source. Tool descriptions are extremely brief (5-10 words) and lack actionable context. No error handling patterns, permission gates, or recovery guidance detected. This cannot be evaluated as an MCP server.
Provide advice through reasoning
Analyse risk for tasks and deadlines
Assign activity to a project
Check upcoming deadlines
Check payment status
Mark a task as completed
Create a new client
Generate a contract
No input schemas visible for any of 35 tools. Cannot verify parameter types, constraints, or validation rules.
| Scored | Grade | Overall | Spec posture | Rubric |
|---|---|---|---|---|
| 2026-09-23 | F | 24 | 2026-07-28+ | v2 |
Create an invoice
Create a new task
Delete a time entry
Draft a client update
Generate a link code
Build timesheet from tracked time
Load activity feed
Summarise activity
Check device status
Check current user focus
Check integration status
Get time summary for user's hours
Gather context from multiple sources
Load clients
Recall saved memories
Load projects
Load tasks
Load time entries
Use the browser to perform an action
Save information to memory
Search for contracts
Send a reminder
Stop the running timer
Sync tasks from external sources
Track time for current activity
Update an existing task
Update a time entry
All tool descriptions are 5-10 words with no context for LLM selection. Descriptions must be 10-1024 chars and explain WHAT, WHEN, and WHY.
No output schemas documented. LLMs cannot plan downstream tool calls or extract required fields for chaining.
Destructive tools (deleteTimeEntry, deleteTask if exists) lack confirmation or dry-run patterns. No recovery guidance in error responses.
No pagination parameters (limit, offset, page) visible for list tools. Large result sets will exhaust context windows.