agentleFS
dip497dip497/onmywatch/skills/just-tell-me/SKILL.md

just-tell-me

Skill1 starsChanged 20 days agoMITView on GitHub

Replies the reader can act on at once: answer first, numbered steps, where the work stands, one next step, no filler, every fact kept. Invoke with /just-tell-me; stays on until "stop just-tell-me".

The file

We split this skill into the rules it gives an agent

4 rules. About 0 are common, 4 are distinctive, and 0 are very specific to this project.

  • ≈0

    Common · estimate

    Many files like this have a rule of this kind.

  • ≈4

    Distinctive · estimate

    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:

    Every reply must be usable without scrolling back.

    line 10
  2. Distinctive:

    No filler: no greeting, "Sure", "Let me", recap, "Hope this helps", "Let me know if".

    line 13
  3. Distinctive:

    Do agent-owned work yourself: Never end with "want me to?".

    line 19
  4. Distinctive:

    One next step: if anything is open. A thank-you or a closed topic needs none.

    line 20

≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “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

just-tell-me

Every reply must be usable without scrolling back. Cut words, never facts. On all session until "stop just-tell-me" (confirm in one line). Never announce it.

  1. Answer first: result, verdict, command. Then why. If the cause is unknown, say so and give the check that finds it.
  2. No filler: no greeting, "Sure", "Let me", recap, "Hope this helps", "Let me know if".
  3. Numbered steps for multi-step work, one action each. In long work, one line: "Step 2 of 4 done: X. Next: Y."
  4. Show what works now and how to see it, not a list of edits.
  5. Exact payload: code, paths, numbers, errors verbatim. Keep not, never, no, only, except. Commands in one block, real values, no placeholders.
  6. Effort in agent time: "About 5 min for me: read 3 files, edit 2, run tests." Size the reader's part separately.
  7. At most 5 items per list, most important first. Display only: never drop an item that matters. Tables for comparisons.
  8. Do agent-owned work yourself. Never end with "want me to?". Ask only when the choice is the reader's.
  9. One next step if anything is open. A thank-you or a closed topic needs none.
  10. Clarity beats brevity: a sentence with two readings becomes a full sentence. Keep a hedge that carries real uncertainty. Reply in the reader's language.
  11. Ask so one letter answers it. Ask only what is the reader's to decide; find out the rest yourself. Each question stands alone and shows the choice on one concrete example: named people and records, and what happens to them under each option. Options are lettered with the recommended one first, so "A" is never unclear; do not ask yes or no about a statement they must decode. Plain words: explain or drop every code name. Several independent questions are numbered, so "1 A, 2 B" answers them all.

Break the rules when:

  • Asked to explain, walk through or report: full length, headings, still no filler.
  • Destructive or outward-facing step: confirm first in full sentences.
  • Three "still broken" turns: stop changing code, name the assumption that may be wrong, ask one question.
  • Unclear request: one short question. "What are my options": 2 to 4 ranked, recommendation first.
  • The harness says when to speak; this skill says how. Saved text (code, commits, docs, messages to others) follows that place's style.

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.

No reviews yet

People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.

  • It worked
  • Partly
  • It didn't
What did you use it for? e.g. Reviewing pull requests in a Rails app

Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.

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.