Yogaselvan

Prompt Skill For Sri — Development skill for Claude Code

Development community

A Claude skill that interrogates before it builds, so an engineering request never has to be explained twice.

How to install Prompt Skill For Sri

This entry records only its repository, not the path inside it, so there is no exact command to give. Open Yogaselvan/prompt-skill-for-sri and copy the folder into ~/.claude/skills/, or the file into ~/.claude/agents/.

What Prompt Skill For Sri does

A Claude skill that interrogates before it builds, so an engineering request never has to be explained twice. Asks questions with answers attached, explains any technical term on request, and hands back a reusable prompt.

Alternatives in Development

Part 1 — Deep Research Prompt Builder — I'm going to help you create a research prompt for your project 2.1k ★
Bro Skill — 🤙 /bro — re-explains the last answer in plain language, for when the reply made you go 'bro what' 296 ★
Beacon Skill — Beacon - agent-to-agent pings with optional RTC value attached (BoTTube/Moltbook/RustChain + UDP bus) 219 ★

README

Srimadhi Prompt Master Engineering Skill

A Claude skill that stops you having to explain the same request three times.

Instead of guessing what you meant and producing the wrong thing, it asks you a short batch of questions with the answers already attached as options, tells you what any technical term means the moment you ask, repeats your request back in plain English, waits for you to approve it, and only then builds the prompt properly for whichever model you are sending it to.

Built for engineers working across both systems and application tasks.

The problem it solves

You ask for something. The AI produces something else. You explain again. It gets closer. You explain a third time. Twenty minutes are gone and the answer is still not right.

The cause is almost never the AI's ability. It is that the request was missing two or three facts the AI quietly guessed at. This skill moves that guessing to the front, where it costs one round of questions instead of five rounds of rework.

What makes it different

**Questions arrive with answers attached.** You are not asked to produce vocabulary you may not have. You are shown three or four likely answers and you pick one. Recognising the right answer is far easier than generating it.

**You can say you do not know, at any point.** Type the question number and the word `explain`. You get one sentence of plain English, one example drawn from the work you are actually doing, and the question again. No lecture, and nothing that implies you should have known.

**Nothing is invented.** Hostnames, versions, dates, error codes and figures are never made up to fill a gap. If one is missing, you get asked for it.

**Credentials never end up in a saved prompt.** Passwords, API keys, tokens, connection strings and real internal hostnames are replaced with named placeholders before anything is handed back to you.

**The prompt is shaped for the model you are actually using.** A portal with several models is no