Looplia e2e

Skill memorysaver/looplia-core/.claude/skills/looplia-e2e

End-to-end testing for looplia CLI from local source. Builds the CLI, initializes workspace, runs the writing-kit workflow, and verifies outputs. Use when testing local development or validating workflow execution.From its SKILL.md

Install

npx -y skills add memorysaver/looplia-core --skill looplia-e2e

Assembled from the repository path, not quoted from the project. Check it against their README if it does not work.

2 things to look at

no licenseNo license file was found in the repository. Code published without one is not open source by default, so using it at work is a question for whoever answers licensing questions where you are.
21 stars21 stars. Stars are a popularity signal and not a quality one, but at this level it is likely that nobody has read this closely except its author, and you would be relying on your own review.

What its file declares

Copied from the file, not written here

The file declares its own license as MIT. That is the author’s claim about this one file, and it is not the same thing as the license GitHub reports for the repository, which is listed with the other numbers below.

SKILL.md

4.3 KB, as published. Nobody here has run it

Looplia E2E Test Skill

End-to-end testing for looplia CLI from local source.

Quick Start

# Set API key in .env
echo "ZENMUX_API_KEY=your-key" >> .env

# Run E2E test
.claude/skills/looplia-e2e/scripts/e2e.sh

Running inside Claude Code? Use env -u CLAUDECODE to allow the CLI to spawn a nested Claude Code subprocess. The CLAUDECODE env var is set by the outer session and blocks any nested claude invocations:
env -u CLAUDECODE .claude/skills/looplia-e2e/scripts/e2e.sh
The scripts internally unset CLAUDECODE in setup_test_env(), so this is only needed if you call e2e.sh directly (not via the looplia-e2e skill invocation).

Modular Test Scripts (v0.8.0)

The E2E tests are modular and can be run individually:

Run All Tests

.claude/skills/looplia-e2e/scripts/e2e.sh

Run Specific Tests

# Auto-discovery only
./scripts/e2e.sh --test auto

# Workflow execution only
./scripts/e2e.sh --test workflow

# Or run directly
./scripts/e2e-auto-discovery.sh

Test Isolation

Tests use LOOPLIA_HOME to isolate from your real workspace:

Test workspace: <project>/test-workspace/
Your workspace: ~/.looplia/ (untouched)

What It Tests

The script performs these steps:

Build - Compiles the CLI from source
Reset - Removes test workspace for fresh start
Init - Initializes workspace with plugins
Configure - Sets provider to ZenMux MiniMax M2.5
Build Command - Tests workflow generation with auto-discovery (HN AI news aggregator)
Run Command - Executes writing-kit workflow with ai-healthcare.md
Verify - Checks outputs, validation state, and auto-discovery results

Expected Outputs

Build command:

test-workspace/workflows/e2e-auto-discovery-test.md    # Generated workflow file
test-workspace/sandbox/build-<id>/
├── validation.json       # workflowValidated: true
└── logs/
    └── *.log             # Execution logs
test-workspace/plugins/auto-discovery-plugin/
├── .claude-plugin/       # Plugin initialized
└── skills/               # Auto-discovered skills

Run command:

test-workspace/sandbox/<run-id>/
├── outputs/
│   ├── summary.json      # Stage 1: Content analysis
│   ├── ideas.json        # Stage 2: Idea generation
│   └── writing-kit.json  # Stage 3: Final output
├── validation.json       # All steps validated: true
└── logs/
    └── *.log             # Execution logs

Success Criteria

Auto-discovery test:

Workflow file created at test-workspace/workflows/e2e-auto-discovery-test.md
validation.json shows workflowValidated: true
Auto-discovery plugin initialized
Skills discovered and cataloged (external dependency)

Workflow test:

3 output files created (summary.json, ideas.json, writing-kit.json)
All 3 steps validated in validation.json
Final output writing-kit.json exists

Troubleshooting

Transient API errors (retry usually works): The ZenMux provider may occasionally return transient errors like "duplicate tool_call id". This is a provider-side issue, not a workflow bug. Simply run the test again:

# Just retry - second run usually succeeds
.claude/skills/looplia-e2e/scripts/e2e.sh

API key issues:

# Verify .env exists and contains ZENMUX_API_KEY
cat .env | grep ZENMUX_API_KEY

Workspace issues:

# Manual reset (test workspace only)
rm -rf test-workspace && looplia init --yes

Build issues:

# Clean rebuild
rm -rf apps/cli/dist && bun run build

File Structure

.claude/skills/looplia-e2e/
├── SKILL.md                    # This file
├── scripts/
│   ├── e2e-setup.sh           # Shared setup/cleanup
│   ├── e2e-auto-discovery.sh  # Auto-discovery focused test
│   └── e2e.sh                 # Orchestrator (runs all tests)
└── assets/
    └── ai-healthcare.md       # Test content fixture