MonkAndMonkey

View saved version →

---
name: monkandmonkey
description: Use MonkAndMonkey tasks, goals, habits, journals, streaks, life areas, tracked time, analytics, and productivity briefs through the hosted MCP.
---

# MonkAndMonkey

Use the MonkAndMonkey MCP as the source of truth for the user's productivity data. Prefer existing summary and analysis tools over rebuilding aggregation in the conversation.

## Daily brief

For a morning or daily brief, request the productivity brief and attention tasks. Keep the result compact and actionable:

1. unfinished work from yesterday and overdue tasks;
2. today's tasks and any active session;
3. habits or streaks at risk;
4. goals, projects, or life areas receiving unusually little attention;
5. at most three concrete next actions.

Use the user's requested date range and timezone. Include journal context only when the user asks for it.

## Reads and analytics

- Use reference-data tools before resolving human-readable areas, projects, tags, or units.
- Use the server's named/custom date ranges and daily or entity grouping instead of manually recomputing totals.
- Distinguish facts returned by MCP from interpretation. Do not invent activity for empty periods.
- Keep large task and journal lists bounded; offer to drill into one area.

## Writes

- Create or update tasks, habits, and journal entries only when the user explicitly asks.
- Use stable record IDs. Resolve an ambiguous name before writing.
- For task completion and planning moves, let the protected API enforce valid state transitions.
- For habit logs, preserve the requested local date, tracking type, unit, and value semantics.
- Summarize exactly what changed after a successful write.

Never perform deletion, arbitrary HTTP requests, direct database access, or session control. If a requested action has no narrow MCP tool, explain the boundary and offer the closest safe supported action.