agentleFS
Sign inStart free
AhmedMerzaAhmedMerza/agent-skills/skills/wait-what/SKILL.md

wait-what

Skill2 starsChanged 41 days agoMITView on GitHub

Stop — that last message did not land. Re-pitch it from a running start, in plain controlled English, keeping the project's domain terms intact. Trigger ONLY on an explicit signal that the user did not follow the previous explanation - /wait-what, "wait, what?", "I don't follow", "you lost me", "say that again". Do NOT fire when "wait, what" signals alarm at an action rather than confusion at an explanation ("wait what, you deleted it?") - answer for the action instead. Adapted from Matt Pocock's wait-what skill (github.com/mattpocock/skills).

We split this skill into the rules it gives an agent

5 rules. 0 are common, 5 are distinctive, and 0 are very specific to this project.

  • 0

    Common

    Most of the 48 edited versions we read keep these.

  • 5

    Distinctive

    This author's own choices. Worth a look.

  • 0

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Wait — I don't understand where you've got to here.

    line 6
  2. Distinctive:

    Don't clarify the confusing sentence.

    line 8
  3. Distinctive:

    Write it in ASD-STE100 Simplified Technical English:

    line 13
  4. Distinctive:

    No jargon, no metaphor, no nested parentheticals.

    line 18
  5. Distinctive:

    If the re-pitch still doesn't land, stop re-pitching and switch modes: ask me questions one at a time until you find where the model in my …

    line 30

“Common” is measured: kept by at least 6 in 10 of the 48 edited versions of this file we could read. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.

Keep what fits and drop the rest. Your version, ready for your agents.
Files
SKILL.md

Wait — I don't understand where you've got to here.

Don't clarify the confusing sentence; that just adds more jargon on top. Re-pitch the whole thing from a running start. Assume the setup is what I missed, not the vocabulary.

Give me a little bit of context first — where we are and why this came up — then the point.

Write it in ASD-STE100 Simplified Technical English:

  • One idea per sentence. Roughly 20 words maximum.
  • Active voice. Present tense where you can.
  • One word, one meaning — don't rotate synonyms for the same thing.
  • No jargon, no metaphor, no nested parentheticals.

Except for this project's own vocabulary — keep those terms exactly as they are. Take them from whichever of CONTEXT.md, CLAUDE.md, the project docs, or the memory index this repo actually has. Simplify the connective tissue around the domain terms, never the terms themselves. Flattening a named setting like max_retry_backoff_ms into "the timing config" makes the re-pitch less useful, not more.

One caution: this register flattens hedges along with jargon. If something was genuinely conditional — "depends on load", "usually but not always" — keep the hedge as its own short sentence rather than dropping it to keep the prose clean.

If the re-pitch still doesn't land, stop re-pitching and switch modes: ask me questions one at a time until you find where the model in my head differs from yours.

Download now (ZIP)

Does it work?

Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.

Reviews can't be read right now.

Your trail

Files you open will collect here, so you can find them again.

Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.

Now do this for your own.

Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.

Check your setup, free