Test Protocol — Turing Pyramid

Quick Start

# Run all tests
./tests/run-tests.sh all

# Run specific suite
./tests/run-tests.sh unit
./tests/run-tests.sh integration
./tests/run-tests.sh regression

When to Run Tests

Required (MUST pass before any release):

After modifying any script in scripts/
After modifying assets/needs-config.json
After modifying assets/cross-need-impact.json
Before publishing to ClawHub

Test Suites

Unit Tests (`tests/unit/`)

Fast, isolated tests for individual functions/formulas.

Test	What it verifies
`test_tension_formula`	tension = importance × (3 - satisfaction)
`test_floor_ceiling`	satisfaction clamped to ≤ 3.0

Integration Tests (`tests/integration/`)

End-to-end tests that run actual scripts.

Test	What it verifies
`test_full_cycle`	run-cycle.sh produces valid output structure

Regression Tests (`tests/regression/`)

Tests for bugs we've found and fixed — prevents re-introduction.

Test	Bug it prevents
`test_socrat_effect`	Floor clamping blocks negative cross-impact at sat=0.5

Adding New Tests

Create tests/<suite>/test_<name>.sh
Make it executable: chmod +x tests/<suite>/test_<name>.sh
Script should exit 0 on pass, exit 1 on fail
Output error message on failure for debugging

Template:

#!/bin/bash
# test_<name>.sh — Description of what this tests

SCRIPT_DIR="$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)"
SKILL_DIR="$(dirname "$(dirname "$SCRIPT_DIR")")"

# Setup
# ...

# Test
# ...

# Verify
if [[ $result == $expected ]]; then
    exit 0
else
    echo "Expected $expected, got $result"
    exit 1
fi

Fixtures

Pre-configured state files in tests/fixtures/:

File	Description
`needs-state-healthy.json`	All needs at sat=2.0-2.5, dep=1
`needs-state-crisis.json`	All needs at sat=0.5, dep=3

Use these for consistent test setup:

cp "$FIXTURES/needs-state-healthy.json" "$STATE_FILE"

Pre-Release Checklist

Before running clawhub publish:

# 1. Run all tests
./tests/run-tests.sh all

# 2. Check for uncommitted changes
git status

# 3. Review diff
git diff HEAD~1

# 4. Update version in SKILL.md
# 5. Commit with descriptive message
# 6. Publish

Tests are not optional. They protect us from ourselves.

2.6 KiB Raw Blame History Unescape Escape