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 takes a name when you run it — a company, a product, a person — and keeps a dedicated document for it: signal-{name}.md. Run it for each name on your list, on demand or on a schedule, and the agent keeps each document up to date independently. Re-running updates the existing doc instead of creating a new one. Good fit when you’re tracking a list — competitors, target accounts, portfolio companies, key people — and you want a tidy, comparable read on each one.

What’s inside

  • Skill: company-signal-tracker (rename it to match what you’re tracking)
  • Tasks or jobs: One per entity, or a single job that runs through a list
  • Knowledge files: signal-{name}.md — one per entity, named automatically
Each document has a quick scorecard at the top, themed sections of named signals, and a short watchlist of things to check on next time.

Set it up

Paste this into the Coconut Assistant. The dimensions are a starting point — swap them for whatever matters to you.
I want a skill that researches a named entity I give it at runtime, and keeps a document at .nut/knowledge/signal-{name}.md. The name should come from the input (lowercase, hyphenated). The filename should always include the entity name — never generic. Group findings into these themes: [customize — e.g. sentiment, AI posture, leadership, tech stack, competitive position]. Each theme contains named signals with a strength rating: Emerging → Building → Established → Fading. Include a scorecard at the top — one row per dimension, with a direction arrow showing the trend. Important: don’t invent sources, and don’t just keep appending. Update existing signals based on fresh evidence. At the bottom, include any readings you considered and rejected (with reasons), plus a watchlist for next time. Once we’ve agreed on the skill, install it on [coconut name]. Then create a task to run it for [first entity].

Running it for more entities

Once the skill is installed, queue up a task for each entity on your list:
Run the company-signal-tracker skill for [entity name]
Each task produces or updates its own signal-{name}.md. The documents are independent, so re-running one doesn’t affect the others.

What the document looks like

# [Entity] — Signal Tracker

> Updated as things change. Signals shift across Emerging → Building → Established → Fading.

## Scorecard
| Dimension | Score | Direction | Rationale |
|---|---|---|---|
| Sentiment | Mixed → trending Positive | ↗ | One-line rationale |
| Perception | Leader | → | One-line rationale |
| Forward-Thinking Index | Early Adopter | ↗ | One-line rationale |

## 1. [Theme name]
**Overall: [one-line read]**

### Named signals
- **[Signal name]***Building.* Why it's at this level, what would move it.

## Considered and rejected
- "[A stronger reading]" — rejected because [evidence].

Keep going

This is the fastest way to track a list of entities side by side. For the bigger picture, see Intelligence Gathering.