deepsource-mcp-server: Command for Claude Code

Analyze GitHub Actions Build

Analyzes a GitHub Actions build to identify failures and their root causes. By default, examines the latest build on the main branch, but you can specify a different run.

Usage

/analyze-build [options]

Options

• --run-id <id>: Analyze a specific workflow run by ID
• --branch <name>: Analyze the latest run from a specific branch (default: main)
• --workflow <name>: Specify which workflow to analyze (default: checks all)
• --pr <number>: Analyze the latest run for a specific PR

Process

1. Identify Target Build

   • Default: Latest workflow run on main branch
   • Or use provided run ID, branch, PR, or workflow name

2. Gather Build Information

   • Workflow run status and conclusion
   • Failed jobs and steps
   • Error logs and annotations
   • Timing information
   • Related commits and PRs

3. Analyze Failures

   • Parse error messages and stack traces
   • Identify failure patterns
   • Categorize error types
   • Determine root causes
   • Check for flaky test indicators

4. Generate Report

   • Executive summary
   • Detailed failure analysis
   • Root cause identification
   • Suggested fixes
   • Historical context (if available)

Analysis Categories

1. Test Failures

• Unit test failures
• Integration test failures
• E2E test failures
• Flaky tests
• Timeout issues

2. Build Errors

• Compilation errors
• TypeScript errors
• Dependency resolution issues
• Missing dependencies
• Version conflicts

3. Linting/Formatting

• ESLint violations
• Prettier formatting issues
• Commit message format
• Markdown linting
• YAML validation

4. Infrastructure Issues

• Runner problems
• Network failures
• API rate limits
• Token/permission issues
• Resource exhaustion

5. Deployment Failures

• Docker build issues
• Registry push failures
• Security scan violations
• Release creation problems

Implementation Steps

# 1. Get the target workflow run
echo "🔍 Identifying target build..."

# Default: latest run on main
if [ -z "$RUN_ID" ]; then
  if [ -n "$PR_NUMBER" ]; then
    # Get latest run for PR
    RUN_ID=$(gh pr checks $PR_NUMBER --json statusCheckRollup --jq '.statusCheckRollup[0].workflowRun.databaseId')
  elif [ -n "$WORKFLOW_NAME" ]; then
    # Get latest run for specific workflow
    RUN_ID=$(gh run list --workflow="$WORKFLOW_NAME" --branch="${BRANCH:-main}" --limit=1 --json databaseId --jq '.[0].databaseId')
  else
    # Get latest run on branch
    RUN_ID=$(gh run list --branch="${BRANCH:-main}" --limit=1 --json databaseId --jq '.[0].databaseId')
  fi
fi

# 2. Fetch run details
echo "📊 Fetching run details for #$RUN_ID..."
gh run view $RUN_ID --json status,conclusion,name,workflowName,event,headBranch,headSha

# 3. Get failed jobs
echo "❌ Identifying failed jobs..."
gh run view $RUN_ID --json jobs --jq '.jobs[] | select(.conclusion == "failure")'

# 4. Fetch logs for failed jobs
echo "📝 Analyzing failure logs..."
gh run view $RUN_ID --log-failed

# 5. Get annotations (errors/warnings)
echo "📌 Checking annotations..."
gh api /repos/{owner}/{repo}/check-runs --jq '.check_runs[].output.annotations[]'