writing-for-agents

Installation
SKILL.md

Every agent-facing artifact wrangles determinism out of a stochastic system. Predictability — the agent taking the same process every run, not producing the same output — is the root virtue; every principle below serves it. A brainstorming artifact should predictably diverge: its tokens vary, its behaviour doesn't.

The principles are universal. The Modes table at the end points to one artifact kind's mechanics each; everything above it holds wherever you're writing.

Leading words — recruit pretrained priors

A leading word (also Leitwort) is a compact concept already in the model's pretraining that the agent thinks with while running the artifact (e.g. lesson, fog of war, tracer bullets). Repeated as a token — never as a sentence — it accumulates a distributed definition and anchors a whole region of behaviour in the fewest tokens, by recruiting priors the model already holds. A strong word may need only one appearance.

It serves predictability twice. In the body it anchors execution: the agent reaches for the same behaviour every time the word appears. In a trigger it anchors invocation: when the same word lives in your prompts, docs, and code, the agent links that shared language to the artifact and fires it more reliably.

Coining your own works if you define it clearly, but a made-up word recruits no priors — you pay in definition tokens what a pretrained word gives free. Reach for an existing word first.

Hunt for refactors to leading words. A triad spelled out at three sites (duplication), a trigger spending a sentence to gesture at one idea — each begs to collapse into a single token:

  • "fast, deterministic, low-overhead" → tight — one quality restated across a phase, into a single pretrained word (a tight loop).
  • "a loop you believe in" → red — a fuzzy gate becomes a binary observable state (the loop goes red, or it doesn't).

You win twice: fewer tokens, and a sharper hook for the agent to hang its thinking on. Assume every artifact carries restatements that leading words retire — go find them.

Installs
5
GitHub Stars
3
First Seen
Jul 7, 2026
writing-for-agents — erikpr1994/materialize