Files
Charon/docs/plans/proof-of-concept/test-backend-coverage.SKILL.md
2026-01-26 19:22:05 +00:00

442 lines
12 KiB
Markdown

---
# agentskills.io specification v1.0
name: "test-backend-coverage"
version: "1.0.0"
description: "Run Go backend tests with coverage analysis and threshold validation (minimum 85%)"
author: "Charon Project"
license: "MIT"
tags:
- "testing"
- "coverage"
- "go"
- "backend"
- "validation"
compatibility:
os:
- "linux"
- "darwin"
shells:
- "bash"
requirements:
- name: "go"
version: ">=1.23"
optional: false
- name: "python3"
version: ">=3.8"
optional: false
environment_variables:
- name: "CHARON_MIN_COVERAGE"
description: "Minimum coverage percentage required (overrides default)"
default: "85"
required: false
- name: "CPM_MIN_COVERAGE"
description: "Alternative name for minimum coverage threshold (legacy)"
default: "85"
required: false
- name: "PERF_MAX_MS_GETSTATUS_P95"
description: "Maximum P95 latency for GetStatus endpoint (ms)"
default: "25ms"
required: false
- name: "PERF_MAX_MS_GETSTATUS_P95_PARALLEL"
description: "Maximum P95 latency for parallel GetStatus calls (ms)"
default: "50ms"
required: false
- name: "PERF_MAX_MS_LISTDECISIONS_P95"
description: "Maximum P95 latency for ListDecisions endpoint (ms)"
default: "75ms"
required: false
parameters:
- name: "verbose"
type: "boolean"
description: "Enable verbose test output"
default: "false"
required: false
outputs:
- name: "coverage.txt"
type: "file"
description: "Go coverage profile in text format"
path: "backend/coverage.txt"
- name: "coverage_summary"
type: "stdout"
description: "Summary of coverage statistics and validation result"
metadata:
category: "test"
subcategory: "coverage"
execution_time: "medium"
risk_level: "low"
ci_cd_safe: true
requires_network: false
idempotent: true
---
# Test Backend Coverage
## Overview
Executes the Go backend test suite with race detection enabled, generates a coverage profile, filters excluded packages, and validates that the total coverage meets or exceeds the configured threshold (default: 85%).
This skill is designed for continuous integration and pre-commit hooks to ensure code quality standards are maintained.
## Prerequisites
- Go 1.23 or higher installed and in PATH
- Python 3.8 or higher installed and in PATH
- Backend dependencies installed (`cd backend && go mod download`)
- Write permissions in `backend/` directory (for coverage.txt)
## Usage
### Basic Usage
Run with default settings (85% minimum coverage):
```bash
cd /path/to/charon
.github/skills/scripts/skill-runner.sh test-backend-coverage
```
### Custom Coverage Threshold
Set a custom minimum coverage percentage:
```bash
export CHARON_MIN_COVERAGE=90
.github/skills/scripts/skill-runner.sh test-backend-coverage
```
### CI/CD Integration
For use in GitHub Actions or other CI/CD pipelines:
```yaml
- name: Run Backend Tests with Coverage
run: .github/skills/scripts/skill-runner.sh test-backend-coverage
env:
CHARON_MIN_COVERAGE: 85
```
### VS Code Task Integration
This skill is integrated as a VS Code task:
```json
{
"label": "Test: Backend with Coverage",
"type": "shell",
"command": ".github/skills/scripts/skill-runner.sh test-backend-coverage",
"group": "test"
}
```
Run via: `Tasks: Run Task``Test: Backend with Coverage`
## Parameters
| Parameter | Type | Required | Default | Description |
|-----------|------|----------|---------|-------------|
| verbose | boolean | No | false | Enable verbose test output (-v flag) |
## Environment Variables
| Variable | Required | Default | Description |
|----------|----------|---------|-------------|
| CHARON_MIN_COVERAGE | No | 85 | Minimum coverage percentage required for success |
| CPM_MIN_COVERAGE | No | 85 | Legacy name for minimum coverage (fallback) |
| PERF_MAX_MS_GETSTATUS_P95 | No | 25ms | Max P95 latency for GetStatus endpoint |
| PERF_MAX_MS_GETSTATUS_P95_PARALLEL | No | 50ms | Max P95 latency for parallel GetStatus |
| PERF_MAX_MS_LISTDECISIONS_P95 | No | 75ms | Max P95 latency for ListDecisions endpoint |
**Note**: Performance thresholds are loosened when running with `-race` flag due to overhead.
## Outputs
### Success Exit Code
- **0**: All tests passed and coverage meets threshold
### Error Exit Codes
- **1**: Coverage below threshold or coverage file generation failed
- **Non-zero**: Tests failed or other error occurred
### Output Files
- **backend/coverage.txt**: Go coverage profile (text format)
- Contains coverage data for all tested packages
- Filtered to exclude main packages and infrastructure code
- Used by `go tool cover` for analysis
### Console Output
The skill outputs:
1. Test execution progress (verbose mode)
2. Coverage filtering status
3. Total coverage percentage summary
4. Coverage validation result (pass/fail)
Example output:
```
Filtering excluded packages from coverage report...
Coverage filtering complete
github.com/Wikid82/charon/backend/internal/api/handlers GetStatus 95.2%
...
total: (statements) 87.4%
Computed coverage: 87.4% (minimum required 85%)
Coverage requirement met
```
## Examples
### Example 1: Basic Execution
Run tests with default settings:
```bash
cd /path/to/charon
.github/skills/scripts/skill-runner.sh test-backend-coverage
```
Expected output:
```
Filtering excluded packages from coverage report...
Coverage filtering complete
total: (statements) 87.4%
Computed coverage: 87.4% (minimum required 85%)
Coverage requirement met
```
### Example 2: Higher Coverage Threshold
Enforce stricter coverage requirement:
```bash
export CHARON_MIN_COVERAGE=90
.github/skills/scripts/skill-runner.sh test-backend-coverage
```
If coverage is below 90%:
```
total: (statements) 87.4%
Computed coverage: 87.4% (minimum required 90%)
Coverage 87.4% is below required 90% (set CHARON_MIN_COVERAGE or CPM_MIN_COVERAGE to override)
```
### Example 3: CI/CD with Verbose Output
Run in GitHub Actions with full test output:
```yaml
- name: Run Backend Tests with Coverage
run: |
export VERBOSE=true
.github/skills/scripts/skill-runner.sh test-backend-coverage
```
### Example 4: Pre-commit Hook
Add to `.git/hooks/pre-commit`:
```bash
#!/usr/bin/env bash
echo "Running backend tests with coverage..."
if ! .github/skills/scripts/skill-runner.sh test-backend-coverage; then
echo "❌ Coverage check failed. Commit aborted."
exit 1
fi
echo "✅ Coverage check passed."
```
## Excluded Packages
The following packages are excluded from coverage analysis as they are entrypoints or infrastructure code that don't benefit from unit tests:
- `github.com/Wikid82/charon/backend/cmd/api` - API server entrypoint
- `github.com/Wikid82/charon/backend/cmd/seed` - Database seeding tool
- `github.com/Wikid82/charon/backend/internal/logger` - Logging infrastructure
- `github.com/Wikid82/charon/backend/internal/metrics` - Metrics infrastructure
- `github.com/Wikid82/charon/backend/internal/trace` - Tracing infrastructure
- `github.com/Wikid82/charon/backend/integration` - Integration test utilities
**Rationale**: These packages are primarily initialization code, external integrations, or test harnesses that are validated through integration tests rather than unit tests.
## Error Handling
### Common Errors and Solutions
#### Error: coverage file not generated by go test
**Cause**: Test execution failed before coverage generation
**Solution**: Review test output for failures; fix failing tests
#### Error: go tool cover failed or timed out after 60 seconds
**Cause**: Corrupted coverage data or memory issues
**Solution**:
1. Clear Go cache: `.github/skills/scripts/skill-runner.sh utility-cache-clear-go`
2. Re-run tests
3. Check available memory
#### Error: Coverage X% is below required Y%
**Cause**: Code coverage does not meet threshold
**Solution**:
1. Add tests for uncovered code paths
2. Review coverage report: `go tool cover -html=backend/coverage.txt`
3. If threshold is too strict, adjust `CHARON_MIN_COVERAGE`
#### Error: Coverage filtering failed or timed out
**Cause**: Large coverage file or sed performance issue
**Solution**: The skill automatically falls back to unfiltered coverage; investigate if this occurs frequently
### Exit Codes Reference
| Exit Code | Meaning | Action |
|-----------|---------|--------|
| 0 | Success | Tests passed, coverage met |
| 1 | Coverage failure | Add tests or adjust threshold |
| Non-zero | Test failure | Fix failing tests |
## Performance Considerations
### Execution Time
- **Fast machines**: ~30-60 seconds
- **CI/CD environments**: ~60-120 seconds
- **With -race flag**: +30% overhead
### Resource Usage
- **CPU**: High during test execution (parallel tests)
- **Memory**: ~500MB peak (race detector overhead)
- **Disk**: ~10MB for coverage.txt
### Optimization Tips
1. Run without `-race` for faster local testing (not recommended for CI/CD)
2. Use `go test -short` to skip long-running tests during development
3. Increase `GOMAXPROCS` for faster parallel test execution
## Related Skills
- [test-backend-unit](./test-backend-unit.SKILL.md) - Fast unit tests without coverage
- [security-check-govulncheck](./security-check-govulncheck.SKILL.md) - Go vulnerability scanning
- [build-check-go](./build-check-go.SKILL.md) - Verify Go build succeeds
- [utility-cache-clear-go](./utility-cache-clear-go.SKILL.md) - Clear Go build cache
## Integration with VS Code Tasks
This skill is integrated as a VS Code task defined in `.vscode/tasks.json`:
```json
{
"label": "Test: Backend with Coverage",
"type": "shell",
"command": ".github/skills/scripts/skill-runner.sh test-backend-coverage",
"group": "test",
"problemMatcher": []
}
```
**To run**:
1. Open Command Palette (`Ctrl+Shift+P` or `Cmd+Shift+P`)
2. Select `Tasks: Run Task`
3. Choose `Test: Backend with Coverage`
## Integration with CI/CD
### GitHub Actions
Reference in `.github/workflows/quality-checks.yml`:
```yaml
jobs:
backend-tests:
runs-on: ubuntu-latest
steps:
- uses: actions/checkout@v4
- uses: actions/setup-go@v5
with:
go-version: '1.23'
- name: Run Backend Tests with Coverage
run: .github/skills/scripts/skill-runner.sh test-backend-coverage
```
### Pre-commit Hook
Integrated via `.pre-commit-config.yaml`:
```yaml
repos:
- repo: local
hooks:
- id: backend-coverage
name: Backend Coverage Check
entry: .github/skills/scripts/skill-runner.sh test-backend-coverage
language: system
pass_filenames: false
```
## Notes
- **Race Detection**: This skill always runs with `-race` flag enabled to detect data races. This adds ~30% overhead but is critical for catching concurrency issues.
- **Coverage Filtering**: Packages excluded from coverage are defined in the script itself (not externally configurable) to maintain consistency across environments.
- **Python Dependency**: The skill uses Python for decimal-precision coverage comparison to avoid floating-point rounding issues in bash.
- **Timeout Protection**: Coverage generation has a 60-second timeout to prevent infinite hangs in CI/CD.
- **Idempotency**: This skill is safe to run multiple times; it cleans up old coverage files automatically.
## Troubleshooting
### Coverage Report Empty or Missing
1. Check that tests exist in `backend/` directory
2. Verify Go modules are downloaded: `cd backend && go mod download`
3. Check file permissions in `backend/` directory
### Tests Hang or Timeout
1. Identify slow tests: `go test -v -timeout 5m ./...`
2. Check for deadlocks in concurrent code
3. Disable race detector temporarily for debugging: `go test -timeout 5m ./...`
### Coverage Threshold Too Strict
If legitimate code cannot reach threshold:
1. Review uncovered lines: `go tool cover -html=backend/coverage.txt`
2. Add test cases for uncovered branches
3. If code is truly untestable (e.g., panic handlers), consider adjusting threshold
## Maintenance
### Updating Excluded Packages
To modify the list of excluded packages:
1. Edit the `EXCLUDE_PACKAGES` array in the script
2. Document the reason for exclusion
3. Test coverage calculation after changes
### Updating Performance Thresholds
To adjust performance assertion thresholds:
1. Update environment variable defaults in frontmatter
2. Document the reason for change in commit message
3. Verify CI/CD passes with new thresholds
---
**Last Updated**: 2025-12-20
**Maintained by**: Charon Project Team
**Source**: `scripts/go-test-coverage.sh`
**Migration Status**: Proof of Concept
**Lines of Code**: ~400 lines (under 500-line target)