Easyskillz QA

easyskillz-qa

Comprehensive QA workflow for easyskillz CLI - build, install locally, and test all flows.

Purpose

Automatically build the easyskillz package, install it in a temporary test environment, and run exhaustive tests covering:

• All command flows (skill, tool, project, docs domains)
• AI-first design principles (one-shot execution, AI detection)
• Error handling and edge cases
• Command design philosophy compliance
• User experience friction points

When to Use

Run this skill whenever:

• You've made changes to easyskillz and want to verify everything works
• Before publishing a new version
• After refactoring or adding new features
• When investigating bug reports
• To validate the entire user experience

Workflow

1. Build Package Locally

# In the easyskillz project root
npm pack

This creates a tarball like easyskillz-2.0.0-alpha.4.tgz in the current directory.

2. Create Temporary Test Environment

# Create temp directory (gitignored)
mkdir -p .qa-temp/test-project
cd .qa-temp/test-project

# Initialize a test project
npm init -y

# Create AI tool config files to simulate real environment
echo "# Test Project" > CLAUDE.md
echo "# Test Project" > AGENTS.md
echo "root = true" > .cursorrules

3. Install Local Package

# Install from the tarball (not from npm)
npm install -g ../../easyskillz-*.tgz

4. Test All Command Flows

A. Project Sync Flow (AI One-Shot)

Test: One-shot execution with all flags

easyskillz project sync --docs=yes --docs-strategy=unified --gitignore=full

Expected:

• ✅ No interactive prompts
• ✅ Detects tools (claude, codex, cursor)
• ✅ Creates .easyskillz/ structure
• ✅ Wires skills via symlinks
• ✅ Centralizes docs to .easyskillz/docs/
• ✅ Updates .gitignore
• ✅ Exit code 0

Verify:

• .easyskillz/easyskillz.json exists with correct config
• .easyskillz/skills/ directory exists
• .easyskillz/docs/INSTRUCTION.md exists (unified strategy)
• .gitignore contains easyskillz entries
• Tool directories (.claude/skills/, .codex/skills/, .cursor/skills/) exist
• Meta-skill easyskillz-reference is present

B. Project Sync Flow (Interactive - Human)

Test: Interactive prompts for humans

# Delete config first
rm .easyskillz/easyskillz.json

# Run without flags (should prompt)
echo -e "y\n1\n1\n" | easyskillz project sync

Expected:

• ✅ Prompts: "Manage instruction files? [Y/n]"
• ✅ Prompts: "Strategy? [1=unified, 2=tool-specific]"
• ✅ Prompts: "Gitignore strategy? [1=full, 2=conflict-only, 3=none]"
• ✅ Same outcome as one-shot

Verify:

• Config saved with user choices
• No crashes or errors

C. Skill Management Flow

Test: Add skill

easyskillz skill add test-skill

Expected:

• ✅ Creates .easyskillz/skills/test-skill/SKILL.md
• ✅ Wires to all registered tools
• ✅ Shows confirmation message

Verify:

• Skill file exists
• Symlinks created in .claude/skills/test-skill/, .codex/skills/test-skill/, .cursor/skills/test-skill/

Test: List skills

easyskillz skill list

Expected:

• ✅ Shows test-skill in active skills
• ✅ Shows meta-skills (easyskillz-reference, etc.)

Test: Deactivate skill

easyskillz skill deactivate test-skill

Expected:

• ✅ Renames to .test-skill.disabled
• ✅ Removes symlinks from tools
• ✅ Shows restore command

Verify:

• .easyskillz/skills/.test-skill.disabled/ exists
• Symlinks removed from tool directories

Test: Activate skill

easyskillz skill activate test-skill

Expected:

• ✅ Restores from .test-skill.disabled to test-skill
• ✅ Re-wires symlinks

Test: Remove skill (AI mode - requires --confirm)

easyskillz skill remove test-skill --confirm

Expected:

• ✅ Permanently deletes skill
• ✅ No prompts (flag provided)
• ✅ Removes all symlinks

Verify:

• Skill directory deleted
• Symlinks removed

D. Tool Management Flow

Test: List tools

easyskillz tool list

Expected:

• ✅ Shows registered tools (claude, codex, cursor)

Test: Register new tool

easyskillz tool register windsurf

Expected:

• ✅ Adds to config
• ✅ Wires all existing skills
• ✅ Updates instruction file
• ✅ Updates .gitignore

Test: Unregister tool (AI mode)

easyskillz tool unregister windsurf --mode=revert --confirm

Expected:

• ✅ Removes from config
• ✅ Keeps tool directory (revert mode)
• ✅ No prompts

Test: Unregister tool (full delete)

# Register again first
easyskillz tool register windsurf
easyskillz tool unregister windsurf --mode=full --confirm

Expected:

• ✅ Removes from config
• ✅ Deletes tool directory

E. Docs Management Flow

Test: List docs

easyskillz docs list

Expected:

• ✅ Shows centralized instruction files
• ✅ Shows strategy (unified or tool-specific)

Test: Sync docs

# Create a new instruction file
echo "# New Doc" > NEW_INSTRUCTION.md

easyskillz docs sync

Expected:

• ✅ Detects new file
• ✅ Centralizes to .easyskillz/docs/

F. JSON Output Flow

Test: All commands with --json flag

easyskillz skill list --json
easyskillz tool list --json
easyskillz docs list --json

Expected:

• ✅ Valid JSON output
• ✅ Contains ok: true field
• ✅ Contains relevant data

Verify:

• Parse JSON successfully
• No extra text outside JSON

G. Error Handling Flow

Test: Missing required arguments

easyskillz skill add
easyskillz skill remove
easyskillz tool register

Expected:

• ✅ Clear error message
• ✅ Shows usage
• ✅ Exit code 1

Test: Invalid flag values

easyskillz project sync --docs=invalid
easyskillz project sync --docs-strategy=invalid
easyskillz project sync --gitignore=invalid

Expected:

• ✅ Clear error message
• ✅ Shows valid options

Test: Missing flags in one-shot mode

easyskillz project sync --docs=yes
# Missing --docs-strategy

Expected:

• ✅ Error: "Missing required flags"
• ✅ Shows which flags are required

Test: Non-existent skill/tool

easyskillz skill remove nonexistent --confirm
easyskillz skill activate nonexistent
easyskillz tool unregister nonexistent --mode=full --confirm

Expected:

• ✅ Clear error: "not found"
• ✅ Exit code 1

H. AI Detection Flow

Test: Simulate AI agent (set environment variable)

# Windsurf detection
WINDSURF_CASCADE_TERMINAL=1 easyskillz skill remove test-skill

Expected:

• ✅ Shows AI warning
• ✅ Explains what to do
• ✅ Shows questions to ask user
• ✅ Shows flag explanations
• ✅ Shows recommended command
• ✅ Exit code 1 (blocked)

Test: AI with flags (should work)

WINDSURF_CASCADE_TERMINAL=1 easyskillz skill add ai-test-skill
WINDSURF_CASCADE_TERMINAL=1 easyskillz skill remove ai-test-skill --confirm

Expected:

• ✅ No AI warning (flags provided)
• ✅ Executes successfully

I. Help System Flow

Test: Top-level help

easyskillz
easyskillz --help

Expected:

• ✅ Shows domain-based structure
• ✅ Lists all domains
• ✅ Shows examples
• ✅ Shows supported tools

Test: Domain help (future enhancement)

easyskillz skill --help
easyskillz tool --help
easyskillz project --help

Expected:

• ✅ Shows domain-specific commands
• ✅ Shows flags for each command

5. Philosophy Compliance Checks

A. AI-First Design

• All interactive prompts have corresponding flags
• One-shot execution works for all commands
• AI detection prevents getting stuck in prompts
• AI warnings are helpful and specific
• Recommended commands are shown

B. Domain-Based Organization

• All commands follow easyskillz <domain> <action> pattern
• No flat commands (old structure)
• Consistent naming across domains

C. Destructive Operations

• skill remove requires --confirm for AI
• tool unregister requires --mode and --confirm for AI
• Safe operations (list, activate, deactivate) work without flags

D. Error Messages

• Clear and actionable
• Show correct usage
• Reference new command structure (not old)

E. Flag Consistency

• Boolean flags use --flag=<value> format
• Confirmation uses --confirm (no value)
• JSON output uses --json

6. Friction Points Check

Look for:

• Confusing error messages
• Unexpected prompts
• Inconsistent behavior
• Missing feedback
• Unclear next steps
• Broken examples in help text
• Old command references in output

7. Report Generation

Create a report file: .qa-temp/qa-report.md

# QA Report - easyskillz v{version}

**Date:** {date}
**Tester:** AI Agent
**Environment:** {OS}, Node {version}

## Summary

- ✅ Passed: X tests
- ❌ Failed: Y tests
- ⚠️  Warnings: Z issues

## Test Results

### Project Sync Flow
- [✅/❌] One-shot execution
- [✅/❌] Interactive mode
- [✅/❌] Config creation
- [✅/❌] Docs centralization
- [✅/❌] Gitignore update

### Skill Management Flow
- [✅/❌] Add skill
- [✅/❌] List skills
- [✅/❌] Deactivate skill
- [✅/❌] Activate skill
- [✅/❌] Remove skill

### Tool Management Flow
- [✅/❌] List tools
- [✅/❌] Register tool
- [✅/❌] Unregister tool (revert)
- [✅/❌] Unregister tool (full)

### Docs Management Flow
- [✅/❌] List docs
- [✅/❌] Sync docs

### Error Handling
- [✅/❌] Missing arguments
- [✅/❌] Invalid flags
- [✅/❌] Non-existent resources

### AI Detection
- [✅/❌] Blocks interactive mode
- [✅/❌] Shows helpful warnings
- [✅/❌] Allows one-shot execution

## Bugs Found

1. **Bug Title**
   - **Severity:** Critical/High/Medium/Low
   - **Command:** `easyskillz ...`
   - **Expected:** ...
   - **Actual:** ...
   - **Steps to Reproduce:** ...

## Philosophy Violations

1. **Violation Title**
   - **Principle:** AI-First/Domain-Based/etc.
   - **Issue:** ...
   - **Recommendation:** ...

## Friction Points

1. **Friction Title**
   - **Where:** Command/Flow
   - **Issue:** ...
   - **Impact:** User confusion/Extra steps/etc.
   - **Suggestion:** ...

## Recommendations

- [ ] Fix critical bugs before release
- [ ] Update documentation for X
- [ ] Improve error message for Y
- [ ] Add missing flag for Z

8. Cleanup

# Uninstall the test package
npm uninstall -g easyskillz

# Return to project root
cd ../..

# Keep .qa-temp for review (it's gitignored)

Success Criteria

✅ All tests pass - No critical bugs ✅ Philosophy compliance - All principles followed ✅ No friction - Smooth user experience ✅ AI-first works - One-shot execution for all flows ✅ Error handling - Clear, helpful messages ✅ Documentation accurate - Examples work as shown

Notes

• The .qa-temp/ directory is gitignored automatically
• Run this skill before every release
• Update this skill when adding new commands
• Test on multiple OS (Windows, macOS, Linux) if possible
• Test with different Node versions if possible

Example Invocation

# From easyskillz project root
easyskillz skill add easyskillz-qa  # If not already added

# Then just ask the AI:
# "Run the easyskillz-qa skill and give me a full report"

The AI will execute all tests, catch bugs, identify philosophy violations, and report friction points.