debugging（调试）

Mode: Cognitive/Prompt-Driven — No standalone utility script; use via agent context.

Systematic Debugging

Overview

Random fixes waste time and create new bugs. Quick patches mask underlying issues.

Core principle: ALWAYS find root cause before attempting fixes. Symptom fixes are failure.

Violating the letter of this process is violating the spirit of debugging.

Iron Laws

1. NEVER propose or implement a fix before completing Phase 1 root cause investigation — a fix without root cause is a guess that will fail or create new bugs.
2. ALWAYS reproduce the bug reliably before debugging — if you can't reproduce it consistently, you're not debugging the real issue.
3. NEVER make more than one change at a time when testing a hypothesis — multiple simultaneous changes make it impossible to determine which change fixed the problem.
4. ALWAYS stop and question the architecture after 3 failed fix attempts — if each fix reveals a new problem, the issue is architectural, not symptomatic.
5. NEVER skip creating a failing test case before implementing the fix — without a test, you cannot verify the fix worked or that it won't regress.

When to Use

When to Use

Use for ANY technical issue:

• Test failures
• Bugs in production
• Unexpected behavior
• Performance problems
• Build failures
• Integration issues

Use this ESPECIALLY when:

• Under time pressure (emergencies make guessing tempting)
• "Just one quick fix" seems obvious
• You've already tried multiple fixes
• Previous fix didn't work
• You don't fully understand the issue

Don't skip when:

• Issue seems simple (simple bugs have root causes too)
• You're in a hurry (rushing guarantees rework)
• Manager wants it fixed NOW (systematic is faster than thrashing)

When to Use: debugging vs smart-debug

Rule of thumb: Start with debugging for straightforward bugs. Escalate to smart-debug when you need hypothesis ranking, structured instrumentation, or the bug is intermittent/production-only.

See also: .claude/skills/smart-debug/SKILL.md

The Four Phases

You MUST complete each phase before proceeding to the next.

Phase 1: Root Cause Investigation

BEFORE attempting ANY fix:

1. Read Error Messages Carefully

- Don't skip past errors or warnings - They often contain the exact solution - Read stack traces completely - Note line numbers, file paths, error codes

1. Reproduce Consistently

- Can you trigger it reliably? - What are the exact steps? - Does it happen every time? - If not reproducible - gather more data, don't guess

1. Check Recent Changes

- What changed that could cause this? - Git diff, recent commits - New dependencies, config changes - Environmental differences

1. Gather Evidence in Multi-Component Systems WHEN system has multiple components (CI - build - signing, API - service - database): BEFORE proposing fixes, add diagnostic instrumentation: For EACH component boundary: - Log what data enters component - Log what data exits component - Verify environment/config propagation - Check state at each layer Run once to gather evidence showing WHERE it breaks THEN analyze evidence to identify failing component THEN investigate that specific component Example (multi-layer system): # Layer 1: Workflow echo "=== Secrets available in workflow: ===" echo "IDENTITY: ${IDENTITY:+SET}${IDENTITY:-UNSET}" # Layer 2: Build script echo "=== Env vars in build script: ===" env | grep IDENTITY || echo "IDENTITY not in environment" # Layer 3: Signing script echo "=== Keychain state: ===" security list-keychains security find-identity -v # Layer 4: Actual signing codesign --sign "$IDENTITY" --verbose=4 "$APP" This reveals: Which layer fails (secrets - workflow OK, workflow - build FAIL) For distributed/microservice systems — prefer OpenTelemetry traces: # Query traces by component (preferred over manual echo/env logging) pnpm trace:query --component <service-name> --event <event-name> --since <ISO-8601> --limit 200 # When trace ID is already known pnpm trace:query --trace-id <traceId> --compact --since <ISO-8601> --limit 200 Fragmented traces (each service has its own root span, trace IDs don't match across boundaries) = broken context propagation. Fix traceparent/tracestate header forwarding before investigating business logic. Instrumentation Gate (before hypothesis generation): If runtime behavior remains unclear after static analysis, add targeted log statements at key decision nodes before generating hypotheses. Use session-scoped log files (.claude/context/tmp/debug-{sessionId}.log) to capture runtime state. Human-in-the-loop: ask the user to reproduce the bug after instrumentation is added, before analyzing results. Only proceed to Phase 2 once runtime evidence is collected.
2. Trace Data Flow WHEN error is deep in call stack: See root-cause-tracing.md in this directory for the complete backward tracing technique. Quick version:

- Where does bad value originate? - What called this with bad value? - Keep tracing up until you find the source - Fix at source, not at symptom

Phase 2: Pattern Analysis

Find the pattern before fixing:

1. Find Working Examples

- Locate similar working code in same codebase - What works that's similar to what's broken?

1. Compare Against References

- If implementing pattern, read reference implementation COMPLETELY - Don't skim - read every line - Understand the pattern fully before applying

1. Identify Differences

- What's different between working and broken? - List every difference, however small - Don't assume "that can't matter"

1. Understand Dependencies

- What other components does this need? - What settings, config, environment? - What assumptions does it make?

Phase 3: Hypothesis and Testing

Scientific method:

1. Form Single Hypothesis

- State clearly: "I think X is the root cause because Y" - Write it down - Be specific, not vague

1. Test Minimally

- Make the SMALLEST possible change to test hypothesis - One variable at a time - Don't fix multiple things at once

1. Verify Before Continuing

- Did it work? Yes - Phase 4 - Didn't work? Form NEW hypothesis - DON'T add more fixes on top

1. When You Don't Know

- Say "I don't understand X" - Don't pretend to know - Ask for help - Research more

Phase 4: Implementation

Fix the root cause, not the symptom:

1. Create Failing Test Case

- Simplest possible reproduction - Automated test if possible - One-off test script if no framework - MUST have before fixing - Use the tdd skill for writing proper failing tests

1. Implement Single Fix

- Address the root cause identified - ONE change at a time - No "while I'm here" improvements - No bundled refactoring

1. Verify Fix

- Test passes now? - No other tests broken? - Issue actually resolved?

1. Cleanup

- Remove all instrumentation added for this debug session (log statements, temporary diagnostics) - Verify cleanup: grep for the session debug ID or instrumentation markers to confirm no debug artifacts remain in production code - Example: rg "debug-{sessionId}" --type-add 'src:*.{js,ts,cjs,mjs}' -tsrc.

1. If Fix Doesn't Work

- STOP - Count: How many fixes have you tried? - If < 3: Return to Phase 1, re-analyze with new information - If >= 3: STOP and question the architecture (step 6 below) - DON'T attempt Fix #4 without architectural discussion

1. If 3+ Fixes Failed: Question Architecture Pattern indicating architectural problem: STOP and question fundamentals: Discuss with your human partner before attempting more fixes This is NOT a failed hypothesis - this is a wrong architecture.

- Each fix reveals new shared state/coupling/problem in different place - Fixes require "massive refactoring" to implement - Each fix creates new symptoms elsewhere - Is this pattern fundamentally sound? - Are we "sticking with it through sheer inertia"? - Should we refactor architecture vs. continue fixing symptoms?

Red Flags - STOP and Follow Process

If you catch yourself thinking:

• "Quick fix for now, investigate later"
• "Just try changing X and see if it works"
• "Add multiple changes, run tests"
• "Skip the test, I'll manually verify"
• "It's probably X, let me fix that"
• "I don't fully understand but this might work"
• "Pattern says X but I'll adapt it differently"
• "Here are the main problems: [lists fixes without investigation]"
• Proposing solutions before tracing data flow
• "One more fix attempt" (when already tried 2+)
• Each fix reveals new problem in different place

ALL of these mean: STOP. Return to Phase 1.

If 3+ fixes failed: Question the architecture (see Phase 4.5)

Your Human Partner's Signals You're Doing It Wrong

Watch for these redirections:

• "Is that not happening?" - You assumed without verifying
• "Will it show us...?" - You should have added evidence gathering
• "Stop guessing" - You're proposing fixes without understanding
• "Ultrathink this" - Question fundamentals, not just symptoms
• "We're stuck?" (frustrated) - Your approach isn't working

When you see these: STOP. Return to Phase 1.

Common Rationalizations

Quick Reference

When Process Reveals "No Root Cause"

If systematic investigation reveals issue is truly environmental, timing-dependent, or external:

1. You've completed the process
2. Document what you investigated
3. Implement appropriate handling (retry, timeout, error message)
4. Add monitoring/logging for future investigation

But: 95% of "no root cause" cases are incomplete investigation.

Supporting Techniques

These techniques are part of systematic debugging and available in this directory:

• root-cause-tracing.md - Trace bugs backward through call stack to find original trigger
• defense-in-depth.md - Add validation at multiple layers after finding root cause
• condition-based-waiting.md - Replace arbitrary timeouts with condition polling
• find-polluter - For test pollution bisection (flaky tests due to shared state): run .claude/tools/analysis/find-polluter/find-polluter.sh (or find-polluter.ps1 on Windows) from the project root to isolate which test pollutes the suite.

Related skills:

• tdd - For creating failing test case (Phase 4, Step 1)
• verification-before-completion - Verify fix worked before claiming success

Real-World Impact

From debugging sessions:

• Systematic approach: 15-30 minutes to fix
• Random fixes approach: 2-3 hours of thrashing
• First-time fix rate: 95% vs 40%
• New bugs introduced: Near zero vs common

AI-Assisted Debugging & Modern Observability (2025+)

OpenTelemetry: The New Stack Trace

For distributed systems, OpenTelemetry traces replace manual echo/env evidence gathering. A trace shows the complete request journey across service boundaries via span IDs and trace IDs (W3C Trace Context standard: traceparent/tracestate headers).

Evidence hierarchy for distributed failures (prefer in order):

1. Distributed traces (OpenTelemetry spans, correlated trace IDs)
2. Structured logs with correlation IDs
3. Metrics with timestamps
4. Manual instrumentation (Phase 1 Step 4 bash examples)

Common symptom — fragmented traces: Each service shows its own root span, trace IDs don't match across boundaries. This means context propagation is broken — fix header forwarding before investigating business logic.

AI-Assisted Root Cause Analysis

LLM-based debugging agents (2025 pattern) augment Phase 1 by reading production traces and correlating with codebase context to suggest minimal reproduction cases.

Use AI assistance for:

• High-complexity distributed failures with multi-service blast radius
• On-call incidents requiring rapid root cause identification
• Converting production traces into deterministic test reproducers

Do NOT skip Phase 1 when using AI assistance. AI suggestions are hypotheses — apply Phase 3 (hypothesis testing) before implementing any AI-suggested fix. AI cannot replace systematic investigation; it accelerates evidence gathering.

Anti-Patterns

Memory Protocol (MANDATORY)

Before starting: Read .claude/context/memory/learnings.md

After completing:

• New pattern -> .claude/context/memory/learnings.md
• Issue found -> .claude/context/memory/issues.md
• Decision made -> .claude/context/memory/decisions.md

ASSUME INTERRUPTION: If it's not in memory, it didn't happen.