linear-continue-debugging

Continue debugging Linear issue: $ARGUMENTS

1. Use mcp__sequential-thinking__sequentialthinking for systematic analysis:

   • Start with thought: "What are all possible causes of this bug?"
   • Generate multiple hypotheses ranked by likelihood
   • Consider edge cases and race conditions
   • Think through implementation details step-by-step
   • Question assumptions and revise thinking as needed
   • Use branch_from_thought for exploring alternative theories

2. Leverage mcp__deepwiki for deep understanding:

   • If bug involves external library: mcp__deepwiki__ask_question(repoName="org/repo", question="How does X handle Y?")
   • Get implementation details: mcp__deepwiki__read_wiki_contents(repoName="org/repo")
   • Understand design decisions that might affect the bug
   • Check for known issues or gotchas in the library

3. Apply scientific debugging:

   • Design specific tests for each hypothesis
   • Add strategic logging with context:

     logger.debug(f"[BUG-{issue_id}] State before: {state}, Inputs: {inputs}, Stack: {inspect.stack()[1].function}")
     

   • Use binary search to isolate problem area
   • Create minimal reproduction case
   • Test in different environments/configurations

4. Advanced debugging techniques:

   • Use debugger with conditional breakpoints
   • Memory profiling if suspecting leaks
   • Performance profiling for timing issues
   • Network inspection for API/communication bugs
   • Database query analysis for data issues

5. Root cause analysis:

   • Use mcp__sequential-thinking to validate findings
   • Distinguish between symptoms and root cause
   • Understand why the bug wasn't caught earlier
   • Identify related code that might have same issue

6. Document and fix:

   • Write failing test that reproduces the bug
   • Implement minimal fix addressing root cause
   • Add defensive coding to prevent recurrence
   • Update Linear with detailed findings:

     mcp__linear__create_comment(issueId, body="""
     ## Root Cause Analysis
     - Hypothesis tested: ...
     - Root cause: ...
     - Why it happened: ...
     - Fix approach: ...
     """)
     

7. Knowledge capture:

   • Create .claude/debugging/[bug-pattern].md with learnings
   • Update documentation if user-facing issue
   • Add code comments explaining non-obvious fixes
   • Consider if architectural change needed to prevent class of bugs

Debugging is detective work. Use all tools available - systematic thinking for hypotheses, deepwiki for understanding, and scientific method for proof.

Take a deep breath in, count 1... 2... 3... and breathe out. You are now centered. Don't hold back. Give it your all.