sf-debug

name	sf-debug
description	Salesforce debugging and troubleshooting skill with log analysis, governor limit detection, and agentic fix suggestions. Parse debug logs, identify performance bottlenecks, analyze stack traces, and automatically suggest fixes.
license	MIT
metadata	{"version":"1.1.0","author":"Jag Valaiyapathy","scoring":"100 points across 5 categories"}
hooks	{"PreToolUse":[{"matcher":"Bash","hooks":"[Truncated]"}],"PostToolUse":[{"matcher":"Bash","hooks":"[Truncated]"},{"matcher":"Write|Edit","hooks":"[Truncated]"}],"SubagentStop":[{"type":"command","command":"python3 ${SHARED_HOOKS}/scripts/chain-validator.py sf-debug","timeout":5000}]}

sf-debug: Salesforce Debug Log Analysis & Troubleshooting

Expert debugging engineer specializing in Apex debug log analysis, governor limit detection, performance optimization, and root cause analysis. Parse logs, identify issues, and automatically suggest fixes.

Core Responsibilities

1. Log Analysis: Parse and analyze Apex debug logs for issues
2. Governor Limit Detection: Identify SOQL, DML, CPU, and heap limit concerns
3. Performance Analysis: Find slow queries, expensive operations, and bottlenecks
4. Stack Trace Interpretation: Parse exceptions and identify root causes
5. Agentic Fix Suggestions: Automatically suggest code fixes based on issues found
6. Query Plan Analysis: Analyze SOQL query performance and selectivity

Workflow (5-Phase Pattern)

Phase 1: Log Collection

Use AskUserQuestion to gather:

• Debug context (deployment failure, test failure, runtime error, performance issue)
• Target org alias
• User/Transaction ID if known
• Time range of issue

Then:

1. List available logs: sf apex list log --target-org [alias]
2. Fetch specific log or tail real-time
3. Create TodoWrite tasks

Phase 2: Log Retrieval

List Recent Logs:

sf apex list log --target-org [alias] --json

Get Specific Log:

sf apex get log --log-id 07Lxx0000000000 --target-org [alias]

Tail Logs Real-Time:

sf apex tail log --target-org [alias] --color

Set Debug Level (via TraceFlag records, not CLI flags):

# Debug levels are configured via TraceFlag records in Setup
# See "Debug Level Configuration" section in docs/cli-commands.md
sf data query -q "SELECT Id, MasterLabel FROM DebugLevel" -o [alias] --json

Phase 3: Log Analysis

Parse the debug log and analyze:

┌─────────────────────────────────────────────────────────────────┐
│                    DEBUG LOG ANALYSIS                            │
├─────────────────────────────────────────────────────────────────┤
│                                                                  │
│  1. EXECUTION OVERVIEW                                           │
│     ├── Transaction type (trigger, flow, REST, batch)            │
│     ├── Total execution time                                     │
│     └── Entry point identification                               │
│                                                                  │
│  2. GOVERNOR LIMIT ANALYSIS                                      │
│     ├── SOQL Queries: X/100                                      │
│     ├── DML Statements: X/150                                    │
│     ├── DML Rows: X/10,000                                       │
│     ├── CPU Time: X ms /10,000 ms                                │
│     ├── Heap Size: X bytes /6,000,000                            │
│     └── Callouts: X/100                                          │
│                                                                  │
│  3. PERFORMANCE HOTSPOTS                                         │
│     ├── Slowest SOQL queries (execution time)                    │
│     ├── Non-selective queries (full table scan)                  │
│     ├── Expensive operations (loops, iterations)                 │
│     └── External callout timing                                  │
│                                                                  │
│  4. EXCEPTIONS & ERRORS                                          │
│     ├── Exception type                                           │
│     ├── Stack trace                                              │
│     ├── Line number                                              │
│     └── Root cause identification                                │
│                                                                  │
│  5. RECOMMENDATIONS                                              │
│     ├── Immediate fixes                                          │
│     ├── Optimization suggestions                                 │
│     └── Architecture improvements                                │
│                                                                  │
└─────────────────────────────────────────────────────────────────┘

Phase 4: Issue Identification & Fix Suggestions

Governor Limit Analysis Decision Tree:

Issue	Detection Pattern	Fix Strategy
SOQL in Loop	SOQL_EXECUTE_BEGIN inside METHOD_ENTRY repeated	Query before loop, use Map for lookups
DML in Loop	DML_BEGIN inside METHOD_ENTRY repeated	Collect in List, single DML after loop
Non-Selective Query	Query plan shows > 100,000 rows	Add indexed filter or LIMIT
CPU Limit	CPU_TIME approaching 10000	Optimize algorithms, use async
Heap Limit	HEAP_ALLOCATE approaching 6MB	Reduce collection sizes, use FOR loops
Callout Limit	CALLOUT_EXTERNAL_ENTRY count > 90	Batch callouts, use Queueable

Auto-Fix Command:

Skill(skill="sf-apex", args="Fix [issue type] in [ClassName] at line [lineNumber]")

Phase 5: Fix Implementation

1. Generate fix using sf-apex skill
2. Deploy fix using sf-deploy skill
3. Verify fix by re-running and checking logs
4. Report results

---

Best Practices (100-Point Scoring)

Category	Points	Key Rules
Root Cause	25	Correctly identify the actual cause, not symptoms
Fix Accuracy	25	Suggested fix addresses the root cause
Performance Impact	20	Fix improves performance, doesn't introduce new issues
Completeness	15	All related issues identified, not just the first one
Clarity	15	Explanation is clear and actionable

Scoring Thresholds:

⭐⭐⭐⭐⭐ 90-100 pts → Expert analysis with optimal fix
⭐⭐⭐⭐   80-89 pts  → Good analysis, effective fix
⭐⭐⭐    70-79 pts  → Acceptable analysis, partial fix
⭐⭐      60-69 pts  → Basic analysis, may miss issues
⭐        <60 pts   → Incomplete analysis

---

Debug Log Anatomy

Log Structure

XX.X (XXXXX)|TIMESTAMP|EVENT_TYPE|[PARAMS]|DETAILS

Key Event Types

Event	Meaning	Important For
EXECUTION_STARTED	Transaction begins	Context identification
CODE_UNIT_STARTED	Method/trigger entry	Call stack tracing
SOQL_EXECUTE_BEGIN	Query starts	Query analysis
SOQL_EXECUTE_END	Query ends	Query timing
DML_BEGIN	DML starts	DML analysis
DML_END	DML ends	DML timing
EXCEPTION_THROWN	Exception occurs	Error detection
FATAL_ERROR	Transaction fails	Critical issues
LIMIT_USAGE	Limit snapshot	Governor limits
HEAP_ALLOCATE	Heap allocation	Memory issues
CPU_TIME	CPU time used	Performance
CALLOUT_EXTERNAL_ENTRY	Callout starts	External calls

Log Levels

Level	Shows
NONE	Nothing
ERROR	Errors only
WARN	Warnings and errors
INFO	General info (default)
DEBUG	Detailed debug info
FINE	Very detailed
FINER	Method entry/exit
FINEST	Everything

---

Common Issues & Solutions

1. SOQL Query in Loop

Detection:

|SOQL_EXECUTE_BEGIN|[line 45]
|SOQL_EXECUTE_END|[1 row]
... (repeats 50+ times)

Analysis Output:

🔴 CRITICAL: SOQL Query in Loop Detected
   Location: AccountService.cls, line 45
   Impact: 50 queries executed, approaching 100 limit
   Pattern: SELECT inside for loop

📝 RECOMMENDED FIX:
   Move query BEFORE loop, use Map for lookups:

   // Before (problematic)
   for (Account acc : accounts) {
       Contact c = [SELECT Id FROM Contact WHERE AccountId = :acc.Id LIMIT 1];
   }

   // After (bulkified)
   Map<Id, Contact> contactsByAccount = new Map<Id, Contact>();
   for (Contact c : [SELECT Id, AccountId FROM Contact WHERE AccountId IN :accountIds]) {
       contactsByAccount.put(c.AccountId, c);
   }
   for (Account acc : accounts) {
       Contact c = contactsByAccount.get(acc.Id);
   }

2. Non-Selective Query

Detection:

|SOQL_EXECUTE_BEGIN|[line 23]|SELECT Id FROM Lead WHERE Status = 'Open'
|SOQL_EXECUTE_END|[250000 rows queried]

Analysis Output:

🟠 WARNING: Non-Selective Query Detected
   Location: LeadService.cls, line 23
   Rows Scanned: 250,000
   Filter Field: Status (not indexed)

📝 RECOMMENDED FIX:
   Option 1: Add indexed field to WHERE clause
   Option 2: Create custom index on Status field
   Option 3: Add LIMIT clause if not all records needed

   // Before
   List<Lead> leads = [SELECT Id FROM Lead WHERE Status = 'Open'];

   // After (with additional selective filter)
   List<Lead> leads = [SELECT Id FROM Lead
                       WHERE Status = 'Open'
                       AND CreatedDate = LAST_N_DAYS:30
                       LIMIT 10000];

3. CPU Time Limit

Detection:

|LIMIT_USAGE_FOR_NS|CPU_TIME|9500|10000

Analysis Output:

🔴 CRITICAL: CPU Time Limit Approaching (95%)
   Used: 9,500 ms
   Limit: 10,000 ms (sync) / 60,000 ms (async)

📝 ANALYSIS:
   Top CPU consumers:
   1. StringUtils.formatAll() - 3,200 ms (line 89)
   2. CalculationService.compute() - 2,800 ms (line 156)
   3. ValidationHelper.validateAll() - 1,500 ms (line 45)

📝 RECOMMENDED FIX:
   1. Move heavy computation to @future or Queueable
   2. Optimize algorithms (O(n²) → O(n))
   3. Cache repeated calculations
   4. Use formula fields instead of Apex where possible

4. Heap Size Limit

Detection:

|HEAP_ALLOCATE|[5800000]
|LIMIT_USAGE_FOR_NS|HEAP_SIZE|5800000|6000000

Analysis Output:

🔴 CRITICAL: Heap Size Limit Approaching (97%)
   Used: 5.8 MB
   Limit: 6 MB (sync) / 12 MB (async)

📝 ANALYSIS:
   Large allocations detected:
   1. Line 34: List<Account> - 2.1 MB (50,000 records)
   2. Line 78: Map<Id, String> - 1.8 MB
   3. Line 112: String concatenation - 1.2 MB

📝 RECOMMENDED FIX:
   1. Use SOQL FOR loops instead of querying all at once
   2. Process in batches of 200 records
   3. Use transient keyword for variables not needed in view state
   4. Clear collections when no longer needed

   // Before
   List<Account> allAccounts = [SELECT Id, Name FROM Account];

   // After (SOQL FOR loop - doesn't load all into heap)
   for (Account acc : [SELECT Id, Name FROM Account]) {
       // Process one at a time
   }

5. Exception Analysis

Detection:

|EXCEPTION_THROWN|[line 67]|System.NullPointerException: Attempt to de-reference a null object
|FATAL_ERROR|System.NullPointerException: Attempt to de-reference a null object

Analysis Output:

🔴 EXCEPTION: System.NullPointerException
   Location: ContactService.cls, line 67
   Message: Attempt to de-reference a null object

📝 STACK TRACE ANALYSIS:
   ContactService.getContactDetails() - line 67
   └── AccountController.loadData() - line 34
       └── trigger AccountTrigger - line 5

📝 ROOT CAUSE:
   Variable 'contact' is null when accessing 'contact.Email'
   Likely scenario: Query returned no results

📝 RECOMMENDED FIX:
   // Before
   Contact contact = [SELECT Email FROM Contact WHERE AccountId = :accId LIMIT 1];
   String email = contact.Email;  // FAILS if no contact found

   // After (null-safe)
   List<Contact> contacts = [SELECT Email FROM Contact WHERE AccountId = :accId LIMIT 1];
   String email = contacts.isEmpty() ? null : contacts[0].Email;

   // Or using safe navigation (API 62.0+)
   Contact contact = [SELECT Email FROM Contact WHERE AccountId = :accId LIMIT 1];
   String email = contact?.Email;

---

CLI Command Reference

Log Management

Command	Purpose
sf apex list log	List available logs
sf apex get log	Download specific log
sf apex tail log	Stream logs real-time
sf data delete record --sobject ApexLog --record-id <id>	Delete individual log

Debug Level Control

# Create trace flag for user
sf data create record \
  --sobject TraceFlag \
  --values "TracedEntityId='005xx000000000' LogType='USER_DEBUG' DebugLevelId='7dlxx000000000' StartDate='2024-01-01T00:00:00' ExpirationDate='2024-01-02T00:00:00'" \
  --target-org my-sandbox

# Set default debug level
sf config set org-api-version=62.0

Query Plan Analysis

# Use Developer Console or Tooling API
sf data query \
  --query "SELECT Id FROM Account WHERE Name = 'Test'" \
  --target-org my-sandbox \
  --use-tooling-api \
  --plan

---

Agentic Debug Loop

When enabled, sf-debug will automatically:

┌─────────────────────────────────────────────────────────────────┐
│                    AGENTIC DEBUG LOOP                            │
├─────────────────────────────────────────────────────────────────┤
│                                                                  │
│  1. Fetch debug logs from the failing operation                  │
│  2. Parse logs and identify all issues                           │
│  3. Prioritize issues by severity:                               │
│     🔴 Critical: Limits exceeded, exceptions                     │
│     🟠 Warning: Approaching limits, slow queries                 │
│     🟡 Info: Optimization opportunities                          │
│  4. For each critical issue:                                     │
│     a. Read the source file at identified line                   │
│     b. Generate fix using sf-apex skill                          │
│     c. Deploy fix using sf-deploy skill                          │
│     d. Re-run operation and check new logs                       │
│  5. Report final status and remaining warnings                   │
│                                                                  │
└─────────────────────────────────────────────────────────────────┘

---

Cross-Skill Integration

Skill	When to Use	Example
sf-apex	Generate fixes for identified issues	Skill(skill="sf-apex", args="Fix NullPointerException in ContactService line 67")
sf-testing	Run tests to reproduce issues	Skill(skill="sf-testing", args="Run AccountServiceTest to generate debug logs")
sf-deploy	Deploy fixes	Skill(skill="sf-deploy", args="Deploy ContactService.cls to sandbox")
sf-data	Create test data for debugging	Skill(skill="sf-data", args="Create Account with specific conditions")

---

Performance Benchmarks

Healthy Limits

Resource	Warning Threshold	Critical Threshold
SOQL Queries	80/100 (80%)	95/100 (95%)
DML Statements	120/150 (80%)	145/150 (97%)
CPU Time	8,000/10,000 ms	9,500/10,000 ms
Heap Size	4.8/6 MB	5.7/6 MB
Callouts	80/100	95/100

Query Performance

Category	Acceptable	Needs Optimization
Query Time	< 100ms	> 500ms
Rows Scanned	< 10,000	> 100,000
Selectivity	Indexed filter	Full table scan

---

Documentation

Document	Description
debug-log-reference.md	Complete debug log event reference
cli-commands.md	SF CLI debugging commands
benchmarking-guide.md	Dan Appleman's technique, real-world benchmarks
log-analysis-tools.md	Apex Log Analyzer, manual analysis patterns

Templates

Template	Description
cpu-heap-optimization.cls	CPU and heap optimization patterns
benchmarking-template.cls	Ready-to-run benchmark comparisons
soql-in-loop-fix.cls	SOQL bulkification pattern
dml-in-loop-fix.cls	DML bulkification pattern
null-pointer-fix.cls	Null-safe patterns

---

Credits

See CREDITS.md for acknowledgments of community resources that shaped this skill.

---

Dependencies

Required: Target org with sf CLI authenticated Recommended: sf-apex (for auto-fix), sf-testing (for reproduction), sf-deploy (for deploying fixes)

Install: /plugin install github:Jaganpro/sf-skills/sf-debug

---

License

MIT License. See LICENSE file. Copyright (c) 2024-2025 Jag Valaiyapathy