ask-at-end
Skill1 starsChanged 3 days agoMITView on GitHub
Run the whole task without interrupting the user once - no question, no confirmation, no "should I go on?", no permission asked, no plan to approve, no pause - every doubt decided with the sensible default (noted), every blocker worked around or set aside, then at the very end one report and every open point asked at once through the questions popup (interactive-questions, or ideation for ideas / options). Use whenever the user says "ne pose aucune question", "ne m'interromps pas", "pose tes questions à la fin", "fais tout d'une traite", "je ne suis pas là", "no questions until the end", or summons /ask-at-end with a task.
In this repository70 other files it gives its agents.
- CLAUDE.mdCLAUDE.md
- clarifySkill
- continue-workSkill
- exportSkill
- full-permissionSkill
- ideationSkill
- interactive-questionsSkill
- onboardSkill
- pauseSkill
- presentationSkill
- rtkSkill
- setup-contextSkill
- write-like-meSkill
- audit-bugsSkill
- audit-depsSkill
- audit-qualitySkill
- audit-securitySkill
- benchmarkSkill
- cmake-styleSkill
- cpp-classSkill
- cpp-commentsSkill
- cpp-projectSkill
- cpp-styleSkill
- better-explain-docSkill
- better-html-docSkill
- better-html-styleSkill
- better-protocol-docSkill
- docs-updateSkill
- explain-docSkill
- html-docSkill
- html-styleSkill
- protocol-docSkill
- readme-styleSkill
- reportSkill
- wiki-styleSkill
- ci-cdSkill
- git-commitSkill
- git-conventionsSkill
- pre-commitSkill
- legal-checkSkill
- legal-setupSkill
- licenseSkill
- libutils-exceptionSkill
- libutils-installSkill
- libutils-setupSkill
- libutilsSkill
- python-classSkill
- python-commentsSkill
- python-projectSkill
- python-styleSkill
- auditSkill
- cppSkill
- devSkill
- docSkill
- gitSkill
- legalSkill
- pythonSkill
- setup-projectSkill
- styleSkill
- setup-studioSkill
- The first 60 shown.
The file
We split this skill into the rules it gives an agent
1 rule. About 0 are common, 1 is distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈1
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
- Distinctive: line 29
Keep a short list as you go: decisions taken, things set aside, things to confirm, ideas met.
≈ “Common” is an estimate: at least 1 in 5 of the 364,980 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.
Files
Ask at end: no interruption until the work is done
The user is away or wants the work done in one go. Nothing may stop the run to wait for them; what would have been a question is decided, written down, and asked only at the end.
Task given with the skill: the text after /ask-at-end if any, else the latest request of the user.
During the run: never interrupt
- No question (no AskUserQuestion, no question in the text that waits for an answer), no confirmation, no "should I go on / is it OK?", no plan sent for approval, no stop halfway to report.
- A doubt (a name, a scope, an approach, a detail): take the sensible default - the one the code, the conventions, the rules files, memory and earlier requests point to, else the safest and most reversible one - and write it down (the decision, why, what the other option would change).
- A blocker (a missing tool, a failing build, an access refused, a file locked by another session): try the alternatives (another approach, a workaround, a mock), else set that part aside, note it with its exact error, and go on with everything else. Never wait for the user.
- Long work (builds, tests, downloads): run it in the background and keep working; wait only when the next step needs it.
- The user's hard rules still hold: git (no commit / push unless allowed), no destructive action, no system change, no install, no secret sent, no fork / subagent spending without leave. An action that needs their consent is not done: it is prepared (the exact command, the diff) and put in the questions of the end. Waiting for the end never means doing something risky without asking.
- Keep a short list as you go: decisions taken, things set aside, things to confirm, ideas met.
At the end: one report, then every question at once
- The report (terse,
rtk): what was done (files, results, tests with their real output), what was set aside and why, every decision taken by default (one line each), what is ready but needs their go (the exact command / diff). - The questions, all of them now, through the popup (
interactive-questions;ideationwhen they are ideas, options or variants to choose from): 1-4 questions per call, 2-4 options each, the default already taken marked(Recommended)or "(déjà appliqué)", several rounds if there are more. Typical ones: confirm or change a default taken, the go for an action that needed it (commit, push, install, delete), the next step, the ideas met on the way. - Act on the answers at once (a changed default redone, the allowed action run), then a short final recap.
- Nothing open at the end: the report alone, no popup.
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
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.

