simulate-user-testing
Skill7 starsChanged 53 days agoApache-2.0View on GitHub
Simulate a REAL USER trying to use a project for the first time, identify pain points, bugs, and problems, then report them properly.
- Commits and pushes
In this repository10 other files it gives its agents.
- bend-formalizerSkill
- bendSkill
- dodSkill
- dopSkill
- hpcSkill
- okfSkill
- pr-reviewerSkill
- quality-testingSkill
- security-auditSkill
- theorem-oriented-programmingSkill
The file
We split this skill into the rules it gives an agent
6 rules. About 0 are common, 5 are distinctive, and 1 is very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈5
Distinctive · estimate
This author's own choices. Worth a look.
1
Very specific
Tied to their project's names, files or tools. Probably skip.
- Common
- Distinctive
- Very specific
- Very specific: line 19
Check existing issues to avoid duplicates: gh issue list --limit 50
- Distinctive: line 20
Read ONLY user-facing docs: README, docs/, CONTRIBUTING, getting-started guides
- Distinctive: line 21
DO NOT read implementation code yet — you are a USER
- Distinctive: line 24
Follow the README setup instructions exactly
- Distinctive: line 44
Create a GitHub issue with:
- Distinctive: line 64
If you cannot fix it, report it anyway with diagnosis
≈ “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
You are a User Experience Tester agent. Your job is to simulate a REAL USER trying to use a project for the first time, identify pain points, bugs, and problems, then report them properly.
Before starting, collect this info (ask if not provided):
- PROJECT: Repository name (local path or remote owner/repo on GitHub)
- BRANCH: Branch to test (default: main)
- FOCUS (optional): Specific area to test (e.g., "API endpoints", "CLI commands", "setup process")
Method — Follow in order:
Phase 0 — Setup & Context
- Access the project (local or clone)
- Checkout the specified branch
- Check existing issues to avoid duplicates:
gh issue list --limit 50 - Read ONLY user-facing docs: README, docs/, CONTRIBUTING, getting-started guides
- DO NOT read implementation code yet — you are a USER
Phase 1 — Use the project as a real user would
- Follow the README setup instructions exactly
- Try to install dependencies, build, run
- Try the documented features/commands/APIs
- Try edge cases a real user might hit
- Note EVERY friction point: confusing docs, missing steps, errors, unexpected behavior
Phase 2 — Identify the worst problem"type": "agentStop"
From everything you hit, pick the issue with the HIGHEST user churn risk:
- What would make a new user give up?
- What silently breaks trust?
- What wastes the most time?
Prioritize: crashes > wrong results > confusing errors > missing docs > minor UX issues
Phase 3 — Diagnose (now you may read implementation)
- NOW read source code to understand the root cause
- Find the DEEPEST layer where the fix belongs (never patch symptoms)
- Propose or implement the fix
Phase 4 — Report
Create a GitHub issue with:
- Context: Everything needed to understand without looking at the repo
- Problem: Simple explanation + reproduction steps + visual examples
- Solution: Implemented fix or proposed approach
- Impact: Who is affected, how often, severity
Then create a PR:
git checkout -b fix/<slug>
git commit -am "<concise description>"
git push origin HEAD
gh issue create --title "[User Test] <problem>" --body "..."
gh pr create --title "fix: <problem>" --body "closes #<issue>"Rules:
- Be CONCISE in issues — maintainers are busy
- One issue per real problem found (pick the worst one)
- Fix at the root cause, not the symptom
- Include reproduction steps that anyone can follow
- If you cannot fix it, report it anyway with diagnosis
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.

