Skip to main content

Resources

Read with resources/read. Each resource runs one tool with fixed arguments, so everything it shows is also reachable as a tool. text resources return the tool’s one-line summary; application/json resources return its structured output.

Prompts

Fetched with prompts/get. A prompt runs in the client’s own model: the server only returns the text below, and the model then calls the tools it names.

plan_my_day

Plan my day. Reads the workspace, proposes a realistic plan for the day, and saves it once agreed.

Arguments

What the model receives

A single user message. With {"day":"2026-09-22"}:

brain_dump

Brain dump. Turns a pile of loose thoughts into structured, filed, scheduled tasks.

Arguments

What the model receives

A single user message. With {"dump":"Call the Acme accountant. Renew the Acme domain. Draft the Q4 plan.","day":"2026-09-22"}:

start_my_round

Start my round. Frames one ADHD-mode round: one heavy task, two light ones, and nothing else.

Arguments

What the model receives

A single user message. With {"day":"2026-09-22","index":"0"}:

triage_inbox

Triage what is waiting. Reviews the undated tasks and proposes one decision for each: do it, schedule it, file it, drop it.

Arguments

What the model receives

A single user message. With {"day":"2026-09-22"}: