diff --git a/.changeset/negation-and-negative-space.md b/.changeset/negation-failure-mode.md similarity index 100% rename from .changeset/negation-and-negative-space.md rename to .changeset/negation-failure-mode.md diff --git a/skills/productivity/writing-great-skills/GLOSSARY.md b/skills/productivity/writing-great-skills/GLOSSARY.md index 59b4137..0269ca8 100644 --- a/skills/productivity/writing-great-skills/GLOSSARY.md +++ b/skills/productivity/writing-great-skills/GLOSSARY.md @@ -164,12 +164,6 @@ _Failure mode._ Steering by prohibition — telling the agent what _not_ to do _Avoid_: ironic rebound, don't-prompting, the pink elephant -### Negative Space - -_Failure mode._ Blindness to the steering done by what you leave _out_. Every decision a skill declines to make, the agent makes from its priors — the unsaid is not neutral, it is delegated. Say nothing about comment length and you have not left length open, you have handed it to the default. The art sense is exact: the shape is carved as much by the space around the subject as by the subject, and here the space is the omission the author never notices leaving. Distinct from **Negation**, which _names_ the wrong thing; Negative Space names nothing and lets the priors choose. Cure: read a draft for the decisions it is silent on, then decide each — fill the space with an explicit target where the default is wrong, and leave it open (a real **branch**) only where any reasonable default will do. - -_Avoid_: the void, omission, silent default - ## Pruning Keeping a skill lean — each remedy paired with the failure it cures. diff --git a/skills/productivity/writing-great-skills/SKILL.md b/skills/productivity/writing-great-skills/SKILL.md index 6250d2a..82abd0d 100644 --- a/skills/productivity/writing-great-skills/SKILL.md +++ b/skills/productivity/writing-great-skills/SKILL.md @@ -81,4 +81,3 @@ Use these to diagnose issues the user may be having with the skill. - **Sprawl** — a skill simply too long, even when every line is live and unique. Hurts readability and maintainability and wastes tokens. The cure is the ladder: disclose **reference** behind pointers, and split by **branch** or sequence so each path carries only what it needs. - **No-op** — a line the model already obeys by default, so you pay load to say nothing. The test: does it change behaviour versus the default? A weak leading word (_be thorough_ when the agent is already thorough-ish) is a no-op; the fix is a stronger word (_relentless_), not a different technique. - **Negation** — steering by prohibition backfires: _don't think of an elephant_ names the elephant and makes it more available, not less. Prompt the **positive** — state the target behaviour so the banned one is never spoken; keep a prohibition only as a hard guardrail you can't phrase positively, and even then pair it with what to do instead. -- **Negative space** — what you leave _out_ steers too: every decision the skill declines, the agent makes from its priors, so the unsaid is delegated, not neutral. Read a draft for what it's silent on, then fill the space with an explicit target where the default is wrong — and leave it open only where any reasonable default does.