skills/development/code-review/SKILL.md
Review ServiceNow code for security vulnerabilities, performance issues, and platform best practices
npx skillsauth add happy-technologies-llc/happy-servicenow-skills code-reviewInstall this skill globally with one command. Works with Claude Code, Cursor, and Windsurf.
3 of 9 scanners reported clean
Some scanners were skipped, did not run, or reported a non-clean status. Review each row below.
This skill covers systematic code review for ServiceNow scripts to identify and remediate:
When to use: Before deploying new scripts to production, during code review processes, when troubleshooting performance issues, or when auditing existing scripts for security compliance.
admin, security_admin, or developer with read access to script tablesdevelopment/code-assist for code generation contextQuery business rules for a specific table:
MCP Approach:
Tool: SN-Query-Table
Parameters:
table_name: sys_script
query: collection=incident^active=true
fields: sys_id,name,collection,when,script,filter_condition,order
limit: 50
REST Alternative:
curl -u "$SN_USER:$SN_PASS" \
"$SN_INSTANCE/api/now/table/sys_script?sysparm_query=collection=incident^active=true&sysparm_fields=sys_id,name,collection,when,script,filter_condition,order&sysparm_limit=50" \
-H "Accept: application/json"
Query script includes:
MCP Approach:
Tool: SN-Query-Table
Parameters:
table_name: sys_script_include
query: active=true^sys_updated_on>=javascript:gs.daysAgo(30)
fields: sys_id,name,script,client_callable,access,api_name
limit: 50
Query client scripts:
MCP Approach:
Tool: SN-Query-Table
Parameters:
table_name: sys_script_client
query: table=incident^active=true
fields: sys_id,name,table,type,fieldname,script,ui_type
limit: 50
Review each script against these critical security patterns:
2a. SQL/GlideRecord Injection:
VULNERABILITY: Unsanitized user input in queries
BAD:
gr.addEncodedQuery(current.variables.user_input);
gr.addQuery('name', request.getParameter('name'));
GOOD:
// Validate and sanitize input
var sanitized = GlideStringUtil.escapeHTML(input);
// Use parameterized queries
gr.addQuery('name', sanitizedName);
// Validate against allowlist
var validFields = ['name', 'number', 'state'];
if (validFields.indexOf(fieldName) === -1) return;
2b. Cross-Site Scripting (XSS):
VULNERABILITY: Unescaped output in UI scripts or HTML fields
BAD:
element.innerHTML = current.short_description;
gs.getMessage(userInput);
GOOD:
element.textContent = current.short_description;
var escaped = GlideStringUtil.escapeHTML(current.short_description);
2c. Privilege Escalation:
VULNERABILITY: Missing role checks or using setWorkflow(false) inappropriately
BAD:
current.setWorkflow(false); // Bypasses business rules and ACLs
current.autoSysFields(false); // Hides audit trail
GOOD:
// Only bypass when absolutely necessary and documented
if (gs.hasRole('admin')) {
current.setWorkflow(false); // REASON: Bulk import requires bypass
}
2d. Sensitive Data Exposure:
VULNERABILITY: Logging or returning sensitive data
BAD:
gs.log('User password: ' + password);
gs.log('SSN: ' + current.social_security_number);
GOOD:
gs.log('Authentication attempted for user: ' + username);
// Never log PII, credentials, or tokens
3a. Unbounded GlideRecord Queries:
ANTI-PATTERN: Query without limit
BAD:
var gr = new GlideRecord('incident');
gr.addQuery('active', true);
gr.query(); // Could return millions of records
GOOD:
var gr = new GlideRecord('incident');
gr.addQuery('active', true);
gr.setLimit(1000); // Always set a reasonable limit
gr.query();
3b. N+1 Query Pattern:
ANTI-PATTERN: Query inside a loop
BAD:
var incidents = new GlideRecord('incident');
incidents.query();
while (incidents.next()) {
var user = new GlideRecord('sys_user');
user.get(incidents.caller_id); // Query per iteration!
// process...
}
GOOD:
// Use dot-walking or join queries
var incidents = new GlideRecord('incident');
incidents.addQuery('active', true);
incidents.query();
while (incidents.next()) {
var callerName = incidents.caller_id.getDisplayValue(); // Dot-walk
// process...
}
// Or use GlideAggregate for counts
var ga = new GlideAggregate('incident');
ga.addQuery('active', true);
ga.groupBy('assignment_group');
ga.addAggregate('COUNT');
ga.query();
3c. Synchronous GlideRecord in Client Scripts:
ANTI-PATTERN: Synchronous server calls from client
BAD:
var gr = new GlideRecord('sys_user');
gr.addQuery('sys_id', userId);
gr.query(); // Synchronous - blocks browser!
if (gr.next()) { ... }
GOOD:
// Use GlideAjax for async server calls
var ga = new GlideAjax('MyScriptInclude');
ga.addParam('sysparm_name', 'getUserInfo');
ga.addParam('sysparm_user_id', userId);
ga.getXMLAnswer(function(answer) {
// Process response asynchronously
});
3d. Unnecessary Field Retrieval:
ANTI-PATTERN: Selecting all fields when only a few are needed
BAD:
var gr = new GlideRecord('incident');
gr.query(); // Returns ALL fields
GOOD:
var gr = new GlideRecord('incident');
gr.addQuery('active', true);
gr.chooseWindow(0, 100);
gr.setFields('sys_id,number,short_description,state');
gr.query();
4a. Deprecated API Usage:
DEPRECATED: Avoid these methods
BAD:
current.getDisplayValue(); // Without field name - ambiguous
gs.log(); // Use gs.info/warn/error instead
Packages.java.lang.*; // Direct Java calls - unsupported
GOOD:
current.getDisplayValue('short_description');
gs.info('Message: {0}', current.number);
4b. Scope Violations:
ANTI-PATTERN: Accessing global scope from scoped app
BAD:
// In scoped app, directly calling global
var result = new IncidentUtils(); // Global script include
GOOD:
// Use scoped API or cross-scope access
var si = new global.IncidentUtils(); // Explicit scope prefix
// Or use sn_ws for REST-based cross-scope communication
4c. Business Rule Ordering:
BEST PRACTICE: Proper ordering
- 100: Data defaults and initialization
- 200: Validation rules
- 300: Data transformation
- 400: Cross-table updates (after rules)
- 500+: Notifications and integrations (after rules)
4d. GlideRecord getValue vs Direct Access:
ANTI-PATTERN: Direct field access for comparisons
BAD:
if (current.state == 6) { ... } // Returns GlideElement, not value
GOOD:
if (current.getValue('state') == '6') { ... }
// Or use typed comparison
if (current.state.changesTo(6)) { ... }
5a. DOM Manipulation:
ANTI-PATTERN: Direct DOM manipulation
BAD:
document.getElementById('incident.short_description').style.color = 'red';
$('incident.state').value = '6';
GOOD:
g_form.setMandatory('short_description', true);
g_form.addDecoration('short_description', 'color_red');
g_form.setValue('state', '6');
5b. Missing isLoading Check:
ANTI-PATTERN: onChange without isLoading guard
BAD:
function onChange(control, oldValue, newValue, isLoading) {
// Fires on every form load, causing unnecessary processing
callServer(newValue);
}
GOOD:
function onChange(control, oldValue, newValue, isLoading) {
if (isLoading || newValue === '') {
return;
}
callServer(newValue);
}
Compile findings into a structured review report:
Code Review Report
================================================
Script: [Script Name]
Table: [Table Name]
Type: [Business Rule / Client Script / Script Include]
Reviewer: [AI-Assisted Review]
Date: [Current Date]
CRITICAL ISSUES:
[C1] Security - SQL Injection in line X
[C2] Security - XSS vulnerability in line Y
HIGH ISSUES:
[H1] Performance - N+1 query pattern in line Z
[H2] Performance - Unbounded GlideRecord query
MEDIUM ISSUES:
[M1] Best Practice - Deprecated API usage
[M2] Best Practice - Missing error handling
LOW ISSUES:
[L1] Maintainability - Missing JSDoc comments
[L2] Maintainability - Variable naming inconsistency
RECOMMENDATIONS:
1. [Specific fix for C1]
2. [Specific fix for C2]
3. [Specific fix for H1]
Update script with fixes:
MCP Approach:
Tool: SN-Update-Record
Parameters:
table_name: sys_script
sys_id: [script_sys_id]
data:
script: |
// [Updated script with fixes applied]
REST Alternative:
curl -u "$SN_USER:$SN_PASS" \
"$SN_INSTANCE/api/now/table/sys_script/[script_sys_id]" \
-H "Content-Type: application/json" \
-X PATCH \
-d '{"script": "// Updated script content"}'
| Action | MCP Tool | REST Endpoint | |--------|----------|---------------| | Query business rules | SN-Query-Table | GET /api/now/table/sys_script | | Query client scripts | SN-Query-Table | GET /api/now/table/sys_script_client | | Query script includes | SN-Query-Table | GET /api/now/table/sys_script_include | | Query UI scripts | SN-Query-Table | GET /api/now/table/sys_ui_script | | Get single script | SN-Get-Record | GET /api/now/table/{table}/{sys_id} | | Update fixed script | SN-Update-Record | PATCH /api/now/table/{table}/{sys_id} |
Symptom: Permission denied when querying sys_script Cause: Missing admin or script-related roles Solution: Verify role assignment:
Tool: SN-Query-Table
Parameters:
table_name: sys_user_has_role
query: user=[user_sys_id]^role.name=admin
fields: sys_id,role,user
Symptom: Script field truncated in API response Cause: API field size limits Solution: Get the specific record directly:
Tool: SN-Get-Record
Parameters:
table_name: sys_script
sys_id: [script_sys_id]
fields: script
Symptom: Need to review all recent changes Solution:
Tool: SN-Query-Table
Parameters:
table_name: sys_script
query: sys_updated_on>=javascript:gs.daysAgo(7)^active=true
fields: sys_id,name,collection,when,sys_updated_by,sys_updated_on
limit: 100
Symptom: Unexpected behavior from multiple rules on same table Solution:
Tool: SN-Query-Table
Parameters:
table_name: sys_script
query: collection=incident^active=true^when=before
fields: sys_id,name,order,filter_condition,action_insert,action_update
limit: 100
# Step 1: Get all active business rules for incident table
Tool: SN-Query-Table
Parameters:
table_name: sys_script
query: collection=incident^active=true
fields: sys_id,name,when,script,order
limit: 100
# Step 2: For each script, check for security patterns
# Look for: addEncodedQuery with user input, setWorkflow(false),
# gs.log with sensitive data, eval() usage, Packages.* calls
# Step 3: Get client scripts for incident
Tool: SN-Query-Table
Parameters:
table_name: sys_script_client
query: table=incident^active=true
fields: sys_id,name,type,script
limit: 50
# Step 4: Check for synchronous GlideRecord in client scripts
# Look for: GlideRecord().query() without callback, direct DOM access
# Get all active script includes with recent updates
Tool: SN-Query-Table
Parameters:
table_name: sys_script_include
query: active=true^sys_updated_on>=javascript:gs.daysAgo(90)
fields: sys_id,name,script,api_name
limit: 100
# For each script include, check for:
# - GlideRecord queries without setLimit()
# - Nested loops with queries (N+1)
# - Missing null checks after .get()
# - Large string concatenation in loops
# Get scripts from the latest update set
Tool: SN-Query-Table
Parameters:
table_name: sys_update_xml
query: update_set=[update_set_sys_id]^nameLIKEsys_script
fields: sys_id,name,action,target_name
limit: 50
# Then retrieve each script for review
Tool: SN-Get-Record
Parameters:
table_name: sys_script
sys_id: [script_sys_id]
fields: name,script,collection,when
development/code-assist - AI-assisted code generationsecurity/acl-management - Access control list managementadmin/update-set-management - Update set handling for deploymentsadmin/instance-hardening - Security hardening guidelinestesting
Manage supplier onboarding, qualification, performance monitoring, and offboarding with auditable lifecycle controls
tools
Identify emerging risks, prioritize intake signals, and route candidates into formal GRC risk assessment workflows
documentation
Screen inbound documents for completeness, policy risk, and routing readiness before extraction or case workflows
testing
Generate concise task summaries with status, timeline, blockers, SLA risk, and recommended next actions