* fix: ContextScout falls back to global core/ when local core/ missing Users who install OAC globally have core standards at ~/.config/opencode/context/core/ but no local core/ directory. ContextScout now does a one-time check on startup: if local core/navigation.md doesn't exist, it checks the global path from paths.json. This resolves the ENOENT errors when agents try to read core/standards/ files that only exist globally. - paths.json: enabled global path (~/.config/opencode/context) - contextscout.md: added global_fallback rule with 2-glob cap - Fallback is ONLY for core/ files, never project-intelligence * docs: add context resolution explanation to README Explains how ContextScout resolves context files: local-first, global fallback only for core/ files, project-intelligence always local. Includes flow diagram, key rules, and common setup table. * fix: resolve installer failures and add comprehensive validation ## Summary Fixes critical installer failures where ContextScout and other components couldn't find their configuration files after installation. Adds comprehensive validation system to prevent future issues. ## Problems Fixed ### 1. Missing Critical Files (13 dead registry references) Files were referenced in registry but didn't exist on disk: - workflows-delegation (split into 3 files) - design-iteration (split into 8 files) - animation-patterns (split into 6 files) - adding-agent (split into 2 files) - adding-skill (split into 3 files) - navigation-design (split into 2 files) - external-libraries (split into 3 files) - claude-code-docs/* (directory removed) ### 2. Missing Infrastructure Files ContextScout failed after installation because 2 critical files weren't in registry: - paths.json - Context root configuration - navigation.md - Root navigation for discovery ### 3. Profile Inconsistencies Profiles referenced components that don't exist: - context:adding-skill → context:adding-skill-basics - context:animation-patterns → animation-basics/components/advanced - context:design-assets (removed - doesn't exist) - context:to-be-consumed/* (removed - directory doesn't exist) - skill:context-manager → subagent:context-manager ## Changes Made ### Registry (registry.json) - Removed 13 dead references - Added 27 new split-file entries - Added 2 critical infrastructure entries - Fixed all 5 profile configurations - Updated ContextScout dependencies - Updated agent-metadata.json dependencies ### New Validation Scripts - scripts/registry/check-dependencies.ts - Comprehensive dependency checker - scripts/tests/test-installer-files.sh - Installer simulation test - scripts/registry/fix-registry.py - Automated registry fixer - scripts/hooks/pre-commit - Git pre-commit hook ### Documentation - docs/maintenance/REGISTRY_FIXES_SUMMARY.md - docs/maintenance/CONTEXTSCOUT_FIX.md - docs/maintenance/REGISTRY_VALIDATION.md ## Validation Results Before: - 13 missing files in registry - 78 orphaned files on disk - 3 missing dependencies - Installer would fail After: - 244 total paths, 0 missing - 262 dependency checks, 0 failures - 0 orphaned dependencies - All profiles consistent - Installer works correctly ## Testing ```bash # Registry validation bun run scripts/registry/validate-registry.ts # ✓ All 244 paths valid # Dependency checking bun run scripts/registry/check-dependencies.ts # ✓ All 262 checks passed # Installer test ./scripts/tests/test-installer-files.sh --local --profile=essential # ✓ All files accessible ``` ## Usage ### Run validation before commits: ```bash bun run scripts/registry/check-dependencies.ts ``` ### Install pre-commit hook: ```bash cp scripts/hooks/pre-commit .git/hooks/pre-commit chmod +x .git/hooks/pre-commit ``` ### Test installer: ```bash ./scripts/tests/test-installer-files.sh --local --profile=essential ``` ## Impact This fix ensures: - ✓ Installer downloads all required files - ✓ ContextScout works immediately after install - ✓ No broken dependencies - ✓ No missing profile components - ✓ Future issues caught by validation ## Breaking Changes None - all changes are additive or fix broken references. ## Related - Fixes: ContextScout failing after installation - Prevents: Missing file errors during install - Adds: Comprehensive validation system * fix: resolve shellcheck warnings in test scripts - Remove unused REPO_URL variable from test-installer-files.sh - Add shellcheck disable comments for cleanup functions used via trap - Fixes CI failures in ShellCheck Analysis workflow * fix: resolve shellcheck errors across all scripts Fixed arithmetic expansion issues and code quality warnings: - Changed ((var++)) to POSIX-compliant 1 syntax - Fixed unquoted variables and grep pipe issues (SC2002, SC2143) - Added shellcheck disable comments for intentional patterns - Fixed variable initialization and return statement placement Scripts fixed: - scripts/tests/*.sh (5 files) - scripts/registry/*.sh (4 files) - scripts/validation/*.sh (2 files) - scripts/*.sh (2 files) - evals/framework/scripts/*.sh (2 files) All scripts now pass shellcheck --severity=warning * chore: trigger CI shellcheck validation |
||
|---|---|---|
| .. | ||
| README.md | ||
| test-collision-detection.sh | ||
| test-compatibility.sh | ||
| test-e2e-install.sh | ||
| test-installer-files.sh | ||
| test-non-interactive.sh | ||
OpenCode Installer Tests
This directory contains test scripts for validating the OpenCode installer functionality and compatibility.
CI Integration: These tests run automatically via .github/workflows/installer-checks.yml on any PR that modifies install.sh, update.sh, or registry.json.
Quick Start
# Run all installer tests
for test in scripts/tests/test-*.sh; do bash "$test"; done
# Run specific test
bash scripts/tests/test-non-interactive.sh
Available Tests
1. Compatibility Test (test-compatibility.sh)
Tests the installer's compatibility across different platforms and bash versions.
Run locally:
bash scripts/tests/test-compatibility.sh
Run remotely:
curl -fsSL https://raw.githubusercontent.com/darrenhinde/OpenAgentsControl/main/scripts/tests/test-compatibility.sh | bash
What it tests:
- ✅ Bash version (3.2+ required)
- ✅ Required dependencies (curl, jq)
- ✅ Script syntax validation
- ✅ Help command functionality
- ✅ List command functionality
- ✅ Profile argument parsing (with and without dashes)
- ✅ Array operations
- ✅ Parameter expansion
- ✅ Temp directory operations
- ✅ File operations
- ✅ Network connectivity
Expected output:
╔════════════════════════════════════════════════════════════════╗
║ OpenCode Installer Compatibility Test ║
╚════════════════════════════════════════════════════════════════╝
System Information:
Platform: macOS
Bash Version: 3.2.57(1)-release
Shell: /bin/bash
[12 tests run...]
╔════════════════════════════════════════════════════════════════╗
║ All Tests Passed! ✓ ║
╚════════════════════════════════════════════════════════════════╝
2. Non-Interactive Mode Test (test-non-interactive.sh)
Critical test - validates piped execution (curl | bash) scenarios.
Run locally:
bash scripts/tests/test-non-interactive.sh
What it tests:
- ✅ Fresh install via pipe (simulates
curl | bash -s profile) - ✅ Collision handling in non-interactive mode
- ✅ All profiles work non-interactively
- ✅ stdin detection and TTY handling
- ✅ NON_INTERACTIVE flag behavior
- ✅ Error handling without user input
Why this matters: Catches bugs where interactive prompts fail silently when stdin is not a terminal.
3. End-to-End Installation Test (test-e2e-install.sh)
Full installation workflow validation.
Run locally:
bash scripts/tests/test-e2e-install.sh
What it tests:
- ✅ Essential profile installs expected files
- ✅ Developer profile installs expected files
- ✅ Custom installation directory works
- ✅ Skip existing files strategy preserves user content
- ✅ File content validity (not empty/corrupted)
- ✅ Directory structure correctness
- ✅ Registry consistency
- ✅ Help and list commands
4. Collision Detection Test (test-collision-detection.sh)
Tests the installer's file collision detection and handling strategies.
Run locally:
bash scripts/tests/test-collision-detection.sh
What it tests:
- ✅ Detection of existing files
- ✅ Skip existing strategy
- ✅ Overwrite strategy
- ✅ Backup & overwrite strategy
- ✅ Collision reporting
- ✅ File grouping by type
Running All Tests
# From repository root
for test in scripts/tests/test-*.sh; do
echo "Running $test..."
bash "$test"
echo ""
done
Test Requirements
All tests require:
- Bash 3.2 or higher
- curl (for network tests)
- jq (for JSON parsing tests)
- Internet connection (for remote fetch tests)
Adding New Tests
When adding new test scripts:
- Naming convention:
test-<feature>.sh - Make executable:
chmod +x scripts/tests/test-<feature>.sh - Add shebang:
#!/usr/bin/env bash - Use colors: Follow existing color scheme (GREEN, RED, YELLOW)
- Exit codes: Exit 0 on success, 1 on failure
- Documentation: Update this README
Example test structure:
#!/usr/bin/env bash
set -e
GREEN='\033[0;32m'
RED='\033[0;31m'
NC='\033[0m'
echo "Testing feature X..."
# Test 1
if [[ condition ]]; then
echo -e "${GREEN}✓${NC} Test 1 passed"
else
echo -e "${RED}✗${NC} Test 1 failed"
exit 1
fi
echo "All tests passed!"
CI/CD Integration
These tests can be integrated into CI/CD pipelines:
GitHub Actions example:
name: Test Installer
on: [push, pull_request]
jobs:
test:
runs-on: ${{ matrix.os }}
strategy:
matrix:
os: [ubuntu-latest, macos-latest]
steps:
- uses: actions/checkout@v2
- name: Install dependencies
run: |
if [[ "$OSTYPE" == "darwin"* ]]; then
brew install jq
else
sudo apt-get install -y jq
fi
- name: Run compatibility tests
run: bash scripts/tests/test-compatibility.sh
Troubleshooting Tests
Test fails with "command not found"
Install missing dependencies:
# macOS
brew install curl jq
# Ubuntu/Debian
sudo apt-get install curl jq
# Fedora/RHEL
sudo dnf install curl jq
Network tests fail
Check internet connectivity:
curl -I https://github.com
Permission errors
Make sure scripts are executable:
chmod +x scripts/tests/*.sh
Test Coverage
Current test coverage:
- ✅ Platform compatibility (macOS, Linux, Windows)
- ✅ Bash version compatibility (3.2+)
- ✅ Argument parsing
- ✅ File collision handling
- ✅ Network operations
- ✅ File system operations
- ✅ Non-interactive/piped execution
- ✅ Profile installation end-to-end
- ✅ Registry validation
- ✅ Error handling scenarios
Future test additions:
- Component dependency resolution
- Rollback functionality
- Windows (Git Bash) specific tests
Contributing
When contributing tests:
- Ensure tests are idempotent (can run multiple times)
- Clean up temporary files/directories
- Provide clear success/failure messages
- Test on multiple platforms if possible
- Update this README with new test documentation