debugging（调试）

Debugging Skill

Debug web applications and Node.js backends using console inspection, network analysis, and non-blocking code debugging with tracepoints, logpoints, and exception monitoring.

When to Use

This skill activates when:

• User reports a bug or error on a web page or backend
• User asks to debug JavaScript (frontend or Node.js)
• User wants to inspect API calls or network requests
• User needs to troubleshoot page loading or API handler issues
• User mentions console errors or warnings
• User wants to debug without breakpoints
• User needs to trace function calls or monitor variables

Capabilities

Console Inspection

browser-devtools-cli o11y get-console-messages
browser-devtools-cli o11y get-console-messages --type warning
browser-devtools-cli --json o11y get-console-messages --type error

Network Analysis

browser-devtools-cli o11y get-http-requests
browser-devtools-cli --json o11y get-http-requests --resource-type fetch
browser-devtools-cli --json o11y get-http-requests --status '{"min":400}'

Tracepoints (Non-Blocking)

browser-devtools-cli debug put-tracepoint --url-pattern "app.js" --line-number 42
browser-devtools-cli debug list-probes --types tracepoint
browser-devtools-cli debug remove-probe --type tracepoint --id <probe-id>
browser-devtools-cli debug clear-probes --types tracepoint

Logpoints

browser-devtools-cli debug put-logpoint --url-pattern "app.js" --line-number 50 --log-expression "user.id"
browser-devtools-cli debug list-probes --types logpoint
browser-devtools-cli debug remove-probe --type logpoint --id <probe-id>
browser-devtools-cli debug clear-probes --types logpoint

Exception Monitoring

browser-devtools-cli debug put-exceptionpoint --state uncaught
browser-devtools-cli debug put-exceptionpoint --state all

Watch Expressions

browser-devtools-cli debug add-watch --expression "this"
browser-devtools-cli debug add-watch --expression "user.id"
browser-devtools-cli debug list-probes --types watch
browser-devtools-cli debug remove-probe --type watch --id <probe-id>
browser-devtools-cli debug clear-probes --types watches

Retrieve and Clear Snapshots (Unified)

browser-devtools-cli --json debug get-probe-snapshots
browser-devtools-cli --json debug get-probe-snapshots --types tracepoint,logpoint,exceptionpoint
browser-devtools-cli --json debug get-probe-snapshots --probe-id "tp_abc" --from-sequence 0
browser-devtools-cli debug clear-probe-snapshots
browser-devtools-cli debug clear-probe-snapshots --types tracepoint --probe-id "tp_abc"

Node.js Backend Debugging (node-devtools-cli)

For debugging Node.js API servers, backends, or scripts:

# 1. Connect to process (by PID, name, or Docker)
node-devtools-cli --session-id backend-debug debug connect --pid 12345
node-devtools-cli --session-id backend-debug debug connect --process-name "server.js"

# 2. Tracepoint on route handler or service
node-devtools-cli --session-id backend-debug debug put-tracepoint \
  --url-pattern "routes/api.ts" \
  --line-number 42

# 3. Exception monitoring
node-devtools-cli --session-id backend-debug debug put-exceptionpoint --state uncaught

# 4. Console logs from Node process
node-devtools-cli --session-id backend-debug --json debug get-logs
node-devtools-cli --session-id backend-debug --json debug get-logs --search "error"

# 5. Retrieve snapshots (after triggering the code path)
node-devtools-cli --session-id backend-debug --json debug get-probe-snapshots
node-devtools-cli --session-id backend-debug --json debug get-probe-snapshots --types tracepoint,exceptionpoint

# 6. Status and cleanup
node-devtools-cli --session-id backend-debug debug status
node-devtools-cli debug disconnect

urlPattern in Node context matches script file paths (e.g., server.js, routes/users.ts). See node-devtools-cli skill for full reference.

Error Investigation

browser-devtools-cli content take-screenshot --name "error-state"
browser-devtools-cli content get-as-html --selector ".error-container"

Basic Debugging Workflow

1. Reproduce: Navigate to the problematic page
2. Capture: Take screenshot of current state
3. Inspect Console: Check for JavaScript errors
4. Analyze Network: Look for failed requests
5. Investigate: Run diagnostic JavaScript
6. Document: Summarize findings with evidence

# Quick debug workflow
browser-devtools-cli navigation go-to --url "https://example.com"
browser-devtools-cli content take-screenshot --name "initial"
browser-devtools-cli --json o11y get-console-messages --type warning
browser-devtools-cli --json o11y get-http-requests --status '{"min":400}'

Advanced Debugging Workflow (Non-Blocking)

1. Set Up Probes

SESSION="--session-id debug-session"

# Navigate to app
browser-devtools-cli $SESSION navigation go-to --url "http://localhost:3000"

# Tracepoint on a function
browser-devtools-cli $SESSION debug put-tracepoint \
  --url-pattern "app.js" \
  --line-number 42

# Exception monitoring
browser-devtools-cli $SESSION debug put-exceptionpoint --state uncaught

2. Add Watch Expressions

browser-devtools-cli $SESSION debug add-watch --expression "this"
browser-devtools-cli $SESSION debug add-watch --expression "user.id"

3. Interact with Application

browser-devtools-cli $SESSION interaction click --selector "#submit-btn"
browser-devtools-cli $SESSION sync wait-for-network-idle

4. Retrieve Snapshots

browser-devtools-cli $SESSION --json debug get-probe-snapshots
browser-devtools-cli $SESSION --json debug get-probe-snapshots --types tracepoint,exceptionpoint
browser-devtools-cli $SESSION --json debug get-probe-snapshots --from-sequence 0

5. Clean Up

browser-devtools-cli $SESSION debug clear-probes
browser-devtools-cli $SESSION debug clear-probe-snapshots
browser-devtools-cli session delete debug-session

Probe Types Summary

Best Practices

1. Choose the right CLI: Use browser-devtools-cli for frontend/page debugging; use node-devtools-cli for Node.js backend/API debugging
2. Always check console for errors first
3. Filter network requests to relevant endpoints
4. Take screenshots before and after actions (browser)
5. Use source maps for minified/bundled code
6. Start with exceptions to catch errors first
7. Use logpoints for lightweight monitoring
8. Poll snapshots with --from-sequence for efficiency
9. Clear probes when done to avoid overhead
10. Document reproduction steps clearly