Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.coconut.dev/llms.txt

Use this file to discover all available pages before exploring further.

​
What you get

A skill that researches upcoming items — events, conferences, product releases, regulatory deadlines — inside a forward-looking window, and keeps a tidy calendar document. Run it again and it updates existing entries with new details instead of duplicating them. Good fit when the answer is a list of dated, real-world items that change over time, and you want one source you can re-run any time.

​
What’s inside

  • Skill: state-marketing-events (rename it for your domain — industry-conferences, product-launches, regulatory-deadlines)
  • Tasks or jobs: One per scope (state, region, category), or a single recurring job
  • Knowledge files: events-marketing-{name}.md — named automatically from the input
Each entry has a clear date, location, type, cost, link, and a short description. Entries are grouped by month, with a summary at the top showing what changed on the latest run.

​
Set it up

Paste this into the Coconut Assistant. The example uses US states; swap the input for your domain.
I want a skill that collates [your item type — e.g. marketing events] for a [your scope — e.g. US state] I give it at runtime, looking ahead [time window — e.g. 6 months]. Save it to .nut/knowledge/events-marketing-{name}.md, where the name comes from the input (lowercase, hyphenated). The filename should always include the scope — never generic. Group entries by month with clear date headings. Each entry should have location, type, cost, link, and a short description. If a date isn’t confirmed, mark it as [Date TBC]. Important: don’t invent items, and don’t duplicate on re-run. Update existing entries only when there’s something new to add. Refresh the “Last updated” line and the summary counts at the top on every run. Once we’ve agreed on the skill, install it on [coconut name]. Then create a task to run it for [first scope value].

​
Running it again

On every run, the agent reads the existing document, compares it against fresh research, and merges. New items get added under the right month. Existing items get updated — a corrected date, an added registration link, a new venue — without being duplicated. Items that fall outside the window roll off.

​
What the document looks like

# [Item Type] — [Scope name]

**Coverage:** May 2026 – November 2026 (6-month window)
**Last updated:** 2026-05-13
**Scope:** [Full scope name]

## Summary
- Total items: 24
- Added this run: 3
- Updated this run: 5
- Months covered: May, Jun, Jul, Aug, Sep, Oct, Nov

## May 2026
### May 14–16, 2026 · [Item name]
| Field | Details |
|---|---|
| Location | [City / venue] |
| Date | May 14–16, 2026 |
| Type | Conference |
| Cost | $1,200 |
| Link | https://… |

About: A short description of the item and why it matters.

​
Keep going

Same pattern as Intelligence Gathering. Taking a name at runtime to drive each run also fits Resource Generation workflows — for example, generating a batch of assets per product or per region.