clarify-before-build
Skill0 starsChanged 30 days agoMITView on GitHub
Use when a task contains ambiguous scope words like "production-ready", "simple", "basic", "quick", "proper", or "clean" — before writing any code.
1 other project uses a version of this.
Adopted by
Most recently taken up by
- jimtin/production-aitheir own version, first seen Jun 11, 2026
Other versions1 project uses a version of this.
Word for word · 0
None found.
Edited · 0
None found.
Wrote their own version · 1
- jimtin/production-ai★ 1same name
How matches are found. Word for word: the exact same text. Edited: shares passages with this file (“99% same” means 99% of its five-word phrases also appear here; 10% or more counts). Their own version: similar meaning, or, weakest, the same name. We can’t tell which project had it first. Projects are counted among the 200 most-starred candidates, and mentions are links to the repository we found, up to 50 per site, so both counts are a floor.
In this repository9 other files it gives its agents.
- AGENTS.mdAGENTS.md
- CLAUDE.mdCLAUDE.md
- .cursor/rules/clean-principles.mdcCursor rule
- architecture-principlesSkill
- clean-codeSkill
- design-patternsSkill
- dry-kiss-yagniSkill
- solid-principlesSkill
- testing-principlesSkill
The file
We split this skill into the rules it gives an agent
3 rules. About 0 are common, 3 are distinctive, and 0 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈3
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 21
Stop and ask when the request contains:
- Distinctive: line 50
Instead of open-ended, offer two concrete options and let them pick:
- Distinctive: line 67
Don't ask when:
≈ “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.
Files
Clarify Before Build
When scope is ambiguous, ask one targeted question before writing a single line.
The Core Problem
Users often say "simple" but mean different things:
- "Simple" = 5 lines, just make it work
- "Simple" = clean architecture, but not over-engineered
- "Simple" = production-ready — they assumed you'd know
Agents often interpret "production-ready" as license to add everything they know belongs in production code. Both assumptions cause rework.
Ambiguous Scope Triggers
Stop and ask when the request contains:
| Trigger word | What's ambiguous |
| "production-ready" | HTML vs plain text? Retry logic? Factory? Logging level? |
| "simple" / "basic" | 10 lines or 100? Error handling included? |
| "quick" | Throwaway prototype or will this live in the codebase? |
| "proper" / "clean" | Their definition, not yours |
| "make it work" | Just the happy path, or error cases too? |
| No qualifier at all | Assume simple until told otherwise |
The Ask
One question. Not five. Pick the one that unlocks the most decisions:
✅ GOOD — one targeted question:
"Before I start: should this be a quick working implementation
or production-grade (error handling, logging, config via env vars)?
The structure will be quite different."
❌ BAD — interrogation:
"What error handling do you want? What about logging?
Should I use SMTP or an email provider? Do you need retries?
What about HTML vs plain text? Should I add a factory?"The Spectrum — Offer Two Endpoints
Instead of open-ended, offer two concrete options and let them pick:
"Quick (20 lines, works for the happy path):
→ csv.reader, returns list of dicts, no error handling
Production-grade (80+ lines):
→ type coercion, encoding detection, row validation,
error reporting, streaming for large files
Which direction?"This is faster than free-form answers and surfaces assumptions immediately.
When NOT to Ask
Don't ask when:
- Task is clearly scoped ("write a one-liner to reverse a string")
- You've already aligned on the approach in this session
- Clarification is more disruptive than just building the simpler version and asking after
Default when unsure: Build the simplest version that works, then explicitly say what you left out and why:
"I kept this simple — no error handling or streaming.
Let me know if you need those."This is faster than asking and delivers something concrete to react to.
Common Agent Mistakes
| Mistake | What agents say | What to do instead |
| Silent scope expansion | "This is production-correct" | Ask first or state what you added and why |
| Assuming "simple" = no quality | Build throwaway code when user wanted clean | Ask or default to clean-but-minimal |
| Over-asking | 5 questions before writing anything | Ask ONE question max |
| Never asking | Implements full stack for a one-liner request | Catch ambiguous triggers above |
After Building — Always State What You Left Out
Even when you don't ask upfront, always close with:
"Left out: retry logic, HTML email, env-var factory.
Add any of these? Just say which."This gives the user control without requiring them to know what to ask for.
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.