agentleFS
MohamedRamadanSalahMohamedRamadanSalah/claude-qa-toolkit/AGENTS.md

claude-qa-toolkit

AGENTS.md0 starsChanged 17 days agoMITView on GitHub

This is a QA/testing toolkit. It contains task instructions ("skills") and Python helper scripts (stdlib only, Python 3.8+, no pip installs needed). Any AI agent working here — not only Claude — must follow these rules. Each skill file contains a mandatory step-by-step workflow. Do not improvise your own format — the workflows exist because their output formats are machine-validated. Run these instead of generating the result yourself: Run each script with --help (or no args) to see usage. After…

  • Installs packages
  • Where files are headingNew AGENTS.md files have more rules about keeping secrets out of the code, and fewer about code style.

    Of the AGENTS.md files we read, the 33,315 first seen in the last six months, against the older ones: the kind of rule that gained most share, and the kind that lost most. About files like this in general, not this one, and an estimate, since a kind is a pattern in the words.

  • In this repository9 other files it gives its agents.

The file

We split this AGENTS.md into the rules it gives an agent

9 rules. About 0 are common, 4 are distinctive, and 5 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.

  • 5

    Very specific

    Tied to their project's names, files or tools. Probably skip.

  • Common
  • Distinctive
  • Very specific
  1. Distinctive:

    Run these instead of generating the result yourself:

    line 26
  2. Very specific:

    Run each script with --help (or no args) to see usage.

    line 33
  3. Distinctive:

    After writing a bug report or test cases, you MUST run the matching validator and fix every ERROR it prints, repeating until it prints OK:

    line 37
  4. Distinctive:

    Never show the user an unvalidated file.

    line 45
  5. Very specific:

    Write outputs to these folders (create if missing): bug-reports/, test-cases/, test-plans/, test-data/, api-tests/, reports/, evidence/.

    line 49
  6. Distinctive:

    Never run tests against production environments without explicit user confirmation.

    line 54
  7. Very specific:

    Never print full auth tokens — mask them (eyJhb...).

    line 55
  8. Very specific:

    No destructive operations (DELETE requests, adb shell pm clear, deleting data) without explicit approval.

    line 56
  9. Very specific:

    Never generate real personal data.

    line 57

≈ “Common” is an estimate: at least 1 in 5 of the 42,283 AGENTS.md 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
AGENTS.md

Instructions for AI agents working in this repository

This is a QA/testing toolkit. It contains task instructions ("skills") and Python helper scripts (stdlib only, Python 3.8+, no pip installs needed). Any AI agent working here — not only Claude — must follow these rules.

Rule 1: When the user asks for a testing task, read the matching skill file FIRST and follow it exactly

User wantsRead this file before doing anything
Test cases / test scenarios from a requirement.claude/skills/test-cases/SKILL.md
Report a bug / write a defect ticket.claude/skills/bug-report/SKILL.md
A test plan or test strategy.claude/skills/test-plan/SKILL.md
Test an API / check endpoints.claude/skills/api-test/SKILL.md
Automate tests (Playwright/Maestro/Appium).claude/skills/automate/SKILL.md
Test data / dummy data files.claude/skills/test-data/SKILL.md
Exploratory testing ideas.claude/skills/exploratory/SKILL.md
Test summary / status report.claude/skills/test-report/SKILL.md
Test the mobile app (Android/iOS).claude/skills/mobile-test/SKILL.md

Each skill file contains a mandatory step-by-step workflow. Do not improvise your own format — the workflows exist because their output formats are machine-validated.

Rule 2: Never do by hand what a script does deterministically

Run these instead of generating the result yourself:

  • python .claude/skills/test-data/scripts/generate_data.py — generates test data files
  • python .claude/skills/test-report/scripts/make_report.py — computes report stats and the release verdict (NEVER calculate percentages or decide go/no-go yourself)
  • python .claude/skills/api-test/scripts/run_api_tests.py — executes API test suites (it blocks production URLs by design; never add --allow-any-env without explicit user confirmation)
  • python .claude/skills/mobile-test/scripts/collect_evidence.py — collects Android bug evidence (device info, screenshot, crash logs) via adb

Run each script with --help (or no args) to see usage.

Rule 3: Validate your own output before showing it to the user

After writing a bug report or test cases, you MUST run the matching validator and fix every ERROR it prints, repeating until it prints OK:

python .claude/skills/bug-report/scripts/validate_bug_report.py <your-file.md>
python .claude/skills/test-cases/scripts/validate_test_cases.py <your-file.md>

Never show the user an unvalidated file. Exit code 0 = acceptable.

Rule 4: File locations

Write outputs to these folders (create if missing): bug-reports/, test-cases/, test-plans/, test-data/, api-tests/, reports/, evidence/.

Rule 5: Safety

  • Never run tests against production environments without explicit user confirmation.
  • Never print full auth tokens — mask them (eyJhb...****).
  • No destructive operations (DELETE requests, adb shell pm clear, deleting data) without explicit approval.
  • Never generate real personal data; use @example.com emails and official test card numbers only.

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.