Error Logger

Overview

This skill maintains a detailed ERROR_LOG.md file that chronicles all errors, bugs, and debugging sessions. Unlike LOG.md (which tracks strategic project evolution), ERROR_LOG.md captures tactical debugging knowledge: what went wrong, why, and how it was fixed.

When to Use This Skill

Use this skill when:

• Encountering an error or exception
• Debugging a tricky issue
• Discovering an edge case
• Solving a technical problem
• Finding a workaround for a limitation
• Identifying a pattern of recurring errors

Typical triggers:

• "Log this error"
• "Document this debugging session"
• "Record how I fixed this bug"

Workflow

Step 1: Identify Error Event

Recognize when to create an error log entry:

Log immediately when:

• Error or exception occurred
• Spent >15 minutes debugging
• Found non-obvious solution
• Discovered edge case
• Implemented workaround

Don't log for:

• Simple typos (unless revealing deeper issue)
• Expected errors (user input validation)
• Trivial fixes (<5 minutes)

Step 2: Gather Error Information

Collect all relevant details:

Required information:

1. Error message: Full traceback or error output
2. Location: File path, line number, function name
3. Context: What were you doing when error occurred?
4. Timestamp: When did this happen?

Optional but helpful: 5. Environment details (Python version, OS, dependencies) 6. Input data that triggered the error 7. Related errors or previous occurrences

How to gather:

# Copy full error traceback
# Note current file and line number
# Describe the action that triggered it

Step 3: Classify Error

Determine error type and severity:

Error Types:

• SYNTAX: Code syntax errors, typos in code
• TYPE: Type errors, None errors, attribute errors
• LOGIC: Logic bugs, incorrect algorithms
• DEPENDENCY: Import errors, version conflicts, missing packages
• ENVIRONMENT: Path issues, permission errors, OS-specific
• PERFORMANCE: Timeouts, memory issues, slow operations
• INTEGRATION: API errors, external service failures
• DATA: Data format issues, parsing errors, validation failures
• EDGE_CASE: Unexpected input or state
• REGRESSION: Previously working code now broken

Severity Levels:

• Critical: Blocks all work, data loss risk
• High: Blocks major functionality
• Medium: Workaround exists, impacts some workflows
• Low: Minor issue, easy to avoid

Step 4: Debug and Find Root Cause

Investigate the underlying issue:

Debugging process:

1. Reproduce the error consistently
2. Isolate the failing component
3. Trace execution flow
4. Identify root cause (not just symptom)
5. Understand why it happened

Document your process:

• What did you try first?
• What led you to the root cause?
• What false leads did you follow?

Root cause should answer:

• Why did the error occur?
• What assumption was violated?
• What condition wasn't handled?

Step 5: Implement Solution

Fix the error and document the solution:

Solution types:

• Fix: Corrected the root cause
• Workaround: Temporary solution, root cause remains
• Refactor: Restructured code to prevent issue
• Config: Changed configuration or environment
• Dependency: Updated/changed dependency version

Document:

• What code changed?
• Why does this fix work?
• Are there any trade-offs?
• Is this temporary or permanent?

Step 6: Add Prevention Strategy

Identify how to avoid this in the future:

Prevention strategies:

• Code pattern to follow (or avoid)
• Validation to add
• Test to write
• Documentation to update
• Design principle to remember

Examples:

• "Always check for None before accessing attributes"
• "Validate input types at function boundaries"
• "Add unit test for empty list edge case"
• "Document API rate limits in code comments"

Step 7: Write Error Log Entry

Create structured entry in ERROR_LOG.md:

Location:

.claude/ERROR_LOG.md

Entry format:

## [ERROR_TYPE] Error Title

**Timestamp:** YYYY-MM-DD HH:MM
**Severity:** [Critical/High/Medium/Low]
**Status:** [Resolved/Workaround/Open]

### Location

- File: `path/to/file.py:123`
- Function: `function_name()`
- Component: [which part of the system]

### Context

[What were you doing when this happened? What was the expected behavior?]

### Error Message

\`\`\`
[Full error traceback or error output]
\`\`\`

### Root Cause

[Why did this error occur? What was the underlying issue?]

### Solution

**Type:** [Fix/Workaround/Refactor/Config/Dependency]

[How was it fixed? What code changed?]

\`\`\`python
# Code snippet showing the fix
[before/after comparison if helpful]
\`\`\`

### Prevention

[How to avoid this in the future?]

- [Prevention strategy 1]
- [Prevention strategy 2]

### Related

- Similar errors: [Links to related error entries]
- Extracted to: [Which skill was updated with this knowledge]
- Issue tracker: [External issue link if applicable]

---

Insertion point:

• New errors go at the top (most recent first)
• Maintains reverse chronological order

Step 8: Link to Knowledge Extraction

Connect error to skill improvements:

If error reveals general knowledge:

1. Note which skill should be updated
2. Add "Extracted to: [skill-name]/references/troubleshooting.md"
3. Later use skill-updater to add troubleshooting guidance

If error is project-specific:

1. Note "Project-specific: stays in ERROR_LOG.md"
2. No extraction needed

Examples:

• Git merge conflict resolution → Extract to git-workflow
• Python type error pattern → Extract to python-best-practices (if exists)
• API rate limiting → Extract to api-integration-skill

Error Log Patterns

Pattern 1: Quick Error

Scenario: Simple error with obvious fix

Example: TypeError from typo

Log format (condensed):

## [TYPE] TypeError: 'NoneType' object is not subscriptable

**Timestamp:** 2025-12-31 14:30
**Severity:** Low
**Status:** Resolved

**Location:** `skill_analyzer.py:45`

**Root Cause:** Forgot to check if section exists before accessing

**Solution:** Added None check: `if section is not None:`

**Prevention:** Always validate before accessing optional data

---

Pattern 2: Complex Debugging Session

Scenario: Multi-hour debugging with false leads

Log format (detailed):

## [PERFORMANCE] Skill loading timeout after 30 seconds

**Timestamp:** 2025-12-31 09:00
**Severity:** High
**Status:** Resolved

**Location:** `.claude/core/skill_loader.py:234`

**Context:**
Loading skills from directory with 50+ skills causes timeout.
Expected load time <1s, actual time >30s.

**Error Message:**
\`\`\`
TimeoutError: Skill loading exceeded 30 second timeout
  at skill_loader.py:234 in load_all_skills()
\`\`\`

**Debugging Process:**
1. First thought: Too many files → profiled file I/O (not the issue)
2. Second attempt: Large SKILL.md files → checked file sizes (all <100KB)
3. Key insight: Noticed repeated regex compilation in loop
4. Root cause found: Regex pattern recompiled for each skill

**Root Cause:**
Regex pattern for parsing frontmatter was compiled inside the loop,
causing O(n) regex compilations instead of O(1).

**Solution:**
**Type:** Refactor

Moved regex compilation outside loop:
\`\`\`python
# Before
for skill in skills:
    pattern = re.compile(r'^---\n(.*?)\n---', re.DOTALL)
    match = pattern.search(skill_content)

# After
pattern = re.compile(r'^---\n(.*?)\n---', re.DOTALL)
for skill in skills:
    match = pattern.search(skill_content)
\`\`\`

Load time reduced from 30s → 0.3s (100x improvement)

**Prevention:**
- Profile before optimizing (avoid premature optimization)
- Move invariant computations out of loops
- Cache compiled regex patterns
- Add performance tests for operations on collections

**Related:**
- Extracted to: skill-creator/references/performance-patterns.md
- Similar pattern: references/common-mistakes.md#loop-invariants

---

Pattern 3: Recurring Error

Scenario: Same error type appears multiple times

Log format (with links):

## [TYPE] AttributeError: 'dict' object has no attribute 'get_value'

**Timestamp:** 2025-12-31 16:45
**Severity:** Medium
**Status:** Resolved

**Context:**
Third occurrence of this pattern - trying to call method on dict
instead of accessing key.

**Root Cause:**
Confusion between dict access (dict['key'] or dict.get('key'))
and object attribute access (obj.attribute).

**Solution:**
Changed to dict access: `config.get('value')` → `config['value']`

**Prevention:**
- Use type hints to catch at development time
- Remember: dicts use [], not .attribute
- Consider using dataclasses for structured config

**Related:**
- Previous occurrences: See [TYPE] entries from 2025-12-30
- Pattern documented in: python-patterns/references/dict-vs-object.md

---

ERROR_LOG.md Management

When to Clean Up

Archive old errors when:

• ERROR_LOG.md exceeds 1000 lines
• Error is >6 months old and resolved
• Knowledge already extracted to skills

Archive location:

• Create

  ERROR_LOG_ARCHIVE_2025.md

• Move old entries
• Keep link in main ERROR_LOG.md

Never delete:

• Recurring error patterns
• Critical errors with complex solutions
• Errors that led to important insights

Integration with knowledge-extractor

knowledge-extractor analyzes ERROR_LOG.md to:

1. Identify recurring error patterns
2. Find troubleshooting knowledge to extract
3. Recommend adding to skill references/troubleshooting.md
4. Generate prevention guidelines

Example extraction:

• 3 git merge conflict errors → Extract pattern to git-workflow
• 5 API timeout errors → Create api-resilience-patterns.md
• 10 None-check errors → Add to python-defensive-programming.md

Resources

references/error_log_template.md

Complete ERROR_LOG.md template with:

• Header structure
• Error type taxonomy
• Example entries for each error type
• Best practices for error documentation

See this file for the full template and examples.

Skill Contract

Stable: 8-step workflow, error classification taxonomy, entry format structure Mutable: Error type categories, example patterns, prevention strategies, integration with knowledge-extractor Update rules: See

references/contract.md

for detailed rules

Full contract specification in

references/contract.md