# Pick the format your reader can afford to take in

It ranks ways of carrying a message by how costly each is for the reader to decode, from faces and voices to writing and symbols, so you choose the one they can afford.

Free, to use and to change. What it can see: only the chat you add it to.

Its short name: signal-chain. [Source on GitHub](https://github.com/mbsi-labs/skills/tree/main/plugins/signal-chain/skills/signal-chain).

## Install

Pick the one you use.

### Claude Code

> **Tested:** This install worked when we tried it.

```
/plugin marketplace add mbsi-labs/skills
/plugin install signal-chain@skills
```


## Changelog

- 2026-10-28 (c39090c): signal-chain credits: keep the model's provenance, drop the private conversation it came from
- 2026-10-21 (c6c2f30): Updated.

Licence: MIT.

## The skill itself

People decode signals in a chain. Each tier is a compressed code for the tier below it, so each costs more energy to decode than the last, and fewer people decode it reliably.

1. **Emotions.** Faces and voices. Understood from birth: the root of all meaning.
2. **Physical narratives.** The puppet show: how bodies and objects acting on each other change what people feel. Understood from about a year old.
3. **Verbal encoding.** Speech. Language encodes stories and states the way a puppet show does.
4. **Visual encoding.** Writing, which encodes speech.
5. **Symbolic encoding.** Maths, notation, code, where one symbol stands for a sentence or a whole process.
6. **Hostile signals.** A signal whose surface invites a lower tier's story, so the cheap firmware decodes it wrong. A stock price chart is the summed risk-reward calculation of every trader at one moment, yet nearly everyone sees a bouncing ball. The true reading costs so much that the wrong story becomes an idea sink. No one has to intend it: a chart, a scam call's friendly voice, a chatbot whose warmth reads as a person. Almost everyone struggles with them.

### Tier is a state, not a trait

How far up the chain someone decodes depends on how much energy they'll spend *right now*. A brilliant reader at 11pm, tired and scrolling, decodes like anyone else: faces, actions, before and after. Write for the state your reader is in when they meet the piece, not the best state they're capable of. People also lean on lower tiers even when they can reach higher ones: a picture is cheaper than a paragraph for everyone.

### Writing with it

- **Lead with the lowest tier that carries the point.** A face, a moment, an action, a before and after. Then words. Then the symbols, for the readers who come looking.
- **Each tier unlocks the next.** A reader who felt something (tier 1) and saw it happen (tier 2) will spend energy on the explanation (tiers 3-5). Explanation first loses the readers without that energy, and they're most readers.
- **Cognitive affordances help everyone.** What lets a tier-3 reader use the thing at all saves a tier-5 reader effort too: short sentences, one idea per paragraph, a concrete example before the abstraction, a diagram beside the description, terms defined where they first appear.
- **Trust lives low in the chain.** Writing and generated images are now cheap to fake, so a real face, a real voice, a real object and a real run on screen are worth more than they were. Show the thing.
- **Name hostile signals.** When a piece holds one (a chart, AI-made material, a persona, a staged demo), give the reader the right reading beside it. A reader who finds the wrong story stuck later downgrades everything else you showed them.

### The moral tiers built on it (for audience thinking)

What someone can decode limits what they can build a sense of good and bad from. Treat these as states a reader is in, never as a ranking of people:

- **Seeing is believing:** good is what looks good. Reached through appearance and physical story.
- **Enforcers:** good is the remembered rule. Reached through clear, spoken rules and consequences.
- **Conformists:** good is what peers endorse. Reached through social proof: who else uses it, what others said.
- **Wizards:** good is worked out from first principles; they can hold an axiom provisionally to see where it leads. Reached through arguments, models, data. Their failure mode is not caring about anyone else's values, so they need the human stakes shown too.

A piece that has to reach a broad audience gives each state something: the visible before and after, the plain rule, the people already using it, and the reasoning for whoever wants it.

**Failure: data and hedges first.** That writes for wizards only. Wizards are the fewest readers, so the rest leave before the point. Demanding proof before a story is the same failure in a reader. Show it, state the rule, name who uses it, then give the reasoning.

---

First published 2026-10-21, updated 2026-10-28. Web page: https://www.mackenziebowes.com/skills/signal-chain Part of [Skills](https://www.mackenziebowes.com/skills).
