Guide · Agent skill
How to use the offensive-sqli agent skill for SQL injection testing
Updated · written with AI from the project's own documentation and checked against our catalogue data
The offensive-sqli skill is a reference playbook that teaches an AI agent how to find and exploit SQL injection flaws in web applications and APIs. It is written for security professionals on authorized penetration tests and bug bounty programs, plus developers and students who want to understand the attacks they must defend against. After this guide you will know what the skill covers, how to install it, how to ask for it in plain language, and what to check before you use it.
SQL Injection Testing PlaybookSnailSploit/Claude-Red · MIT Download the skill Quick start
- 1Get the skill. Download the offensive-sqli ZIP from this page. The skill's main file is SKILL.md.
- 2Install it. Unzip it into ~/.claude/skills/ for personal use or into a project's .claude/skills/ folder, or upload it under Settings → Capabilities → Skills in the Claude apps.
- 3Install the tools you need. Install sqlmap, Burp Suite and any other referenced tools yourself, since they are not bundled with the skill.
- 4Confirm authorization and scope. Write down the hosts, parameters and actions you are permitted to test before you send any request.
- 5Ask in plain language. Describe the task, the input, and the authorized target, for example asking for detection on a specific parameter.
- 6Verify and document. Check every suspected injection by hand, then ask the agent to write up the finding with remediation steps.
What the skill makes the agent do
An agent skill is a folder containing a SKILL.md file with instructions that an AI agent loads when a task matches the skill's description. This one holds a SQL injection testing methodology. When you ask about SQL injection testing, the agent follows a structured workflow: map every input that reaches the database, probe for injection, identify the database type, enumerate the schema, extract data where in scope, and document findings with remediation.
The playbook covers error-based, UNION-based, boolean and time-based blind, out-of-band, second-order, NoSQL, GraphQL, WebSocket and JSON-operator SQL injection. It also includes WAF bypass techniques, database-specific notes for MySQL, MSSQL, PostgreSQL and Oracle, cloud-specific attack paths, ORM CVE tracking, and commands for SQLmap and related tools.
- Map inputs: URL parameters, POST bodies, cookies, headers, API filters and WebSocket messages
- Detect: basic probes, error-based checks, and blind techniques when no error is visible
- Identify the database and enumerate tables and columns
- Exploit within scope, then write up findings and remediation
Install the offensive-sqli skill
There are two ways to install it. For Claude Code, download the ZIP from the skill's page on this directory and unzip it into your user skills folder, ~/.claude/skills/, or into a project's .claude/skills/ folder if you want the skill limited to one project. The skill's files are in the Skills/web/offensive-sqli folder of the SnailSploit/Claude-Red repository, and the listing shows SKILL.md as the file included.
In the Claude apps, you can instead upload the skill under Settings → Capabilities → Skills. Either way, the agent should pick it up when a matching request comes in.
- Claude Code, personal: unzip into ~/.claude/skills/
- Claude Code, one project: unzip into the project's .claude/skills/
- Claude apps: Settings → Capabilities → Skills, then upload
Trigger the skill with natural requests
You do not need a special command. The skill's description says it applies when you are performing web application SQL injection testing, database enumeration, privilege escalation via SQLi, or assessing injection vectors in APIs and modern stacks. Describe the task in those terms and state that the target is authorized.
Good requests name the target, the input you want checked, and the scope. Vague requests such as 'hack this site' are outside what the skill is for, and you should not expect useful help with them.
- Include the authorization and scope: which host, which parameters, and what is off limits
- Name the input type: a URL parameter, a JSON API filter, a WebSocket message, or a login form
- Say what stage you need: detection, database fingerprinting, enumeration, or write-up
- Ask for remediation advice in the same request if you want it in the report
Example requests and what to expect
Here are the kinds of requests the skill is built for. The agent will work through the workflow in the playbook and explain its reasoning, but it still depends on your tools and your authorization being in place.
For an authorized test, a request like 'On my authorized test target, check the id parameter on /page for SQL injection and tell me the database type' fits the detection stage. A request like 'Help me build a SQLmap command to enumerate databases from this saved Burp request file' fits the automation section, which includes a workflow that exports a Burp request to a file and runs SQLmap against it.
- Detection: 'Map the input vectors on this API and suggest which ones to probe first'
- Blind SQLi: 'The app shows no errors. Suggest boolean-based checks for the search parameter'
- Automation: 'Show me how to run sqlmap against this exported request file with --dbs'
- Reporting: 'Write up this confirmed SQLi finding with remediation steps for the developers'
sqlmap -r req.txt --dbsWhat you need before you start
The skill is a reference, not a toolkit. It does not bundle the tools it refers to, so you must install them yourself. The playbook names sqlmap, Burp Suite, ghauri, sublist3r, httpx, waybackurls, gf, hakrawler and arjun. Install only the ones your test needs, and confirm each one runs on your machine before you rely on it.
You also need target systems you are authorized to test, such as your own lab, a deliberately vulnerable practice application, or a bug bounty program's in-scope assets. The listing says the skill runs no scripts and makes no network calls of its own. The tools it points you to do make network requests, so the risk lies in how you use those tools.
- Required and not bundled: sqlmap, Burp Suite, ghauri, sublist3r, httpx, waybackurls, gf, hakrawler, arjun
- Target systems you own or have written permission to test
- A written scope and rollback plan for anything beyond a lab
Safety notes on scripts, network and destructive payloads
The skill itself runs no scripts and uses no network, according to its listing. Its content, however, includes payloads that can damage data or availability if pointed at a live system. These include time-based sleep payloads, DROP TABLE examples, file-write and web shell examples, and OS command execution examples for MSSQL, PostgreSQL and MySQL.
Treat those examples as things to understand, not things to run. Do not use them against production systems without an agreed scope and rollback plan. Exploitation of systems you do not own or lack permission to test is illegal.
- Use only on systems you own or are authorized in writing to test
- Avoid time-based sleep payloads on production; they can slow or disrupt service
- Do not run destructive statements or write files on a live system
- Keep a record of what you tested and what you changed
Tips for getting useful results
Give the agent the exact request you captured, including headers and body, rather than a summary. Burp Suite exports make it easier for SQLmap to reproduce the request. When the application shows no visible errors, ask for boolean-based or time-based checks, and be aware that the playbook notes Ghauri as faster for time-based cases.
Check every finding by hand before you report it. Automated tools produce false positives, and the playbook's own workflow ends with manual verification.
- Paste the full raw request rather than a description
- Ask for boolean-based checks first when errors are hidden
- Verify each automated finding manually before writing it up
- Ask the agent to structure the write-up with remediation
What people use it for
- Running an authorized penetration test and working through input detection, database identification and enumeration in order
- Preparing a bug bounty report on an in-scope API by checking JSON filters and WebSocket messages for injection
- Helping a developer or AppSec engineer understand the attack patterns in ORM and NoSQL code they need to defend
- Building a repeatable SQLmap command sequence from a captured Burp request in a lab environment
- Working through SQL injection concepts in a deliberately vulnerable practice application as a student
Frequently asked questions
How do I install the offensive-sqli skill?
Download the ZIP from this page and unzip it into ~/.claude/skills/ for personal use, or into a project's .claude/skills/ folder. In the Claude apps you can upload it under Settings → Capabilities → Skills instead.
Is the offensive-sqli skill free to use?
The skill is listed under the MIT licence. Check the licence file in the repository for the exact terms before you redistribute it. The tools it references, such as sqlmap and Burp Suite, have their own licences that you must check separately.
What do I need to install alongside the skill?
The skill does not bundle its tools. The playbook refers to sqlmap, Burp Suite, ghauri, sublist3r, httpx, waybackurls, gf, hakrawler and arjun, which you must install yourself. You also need a target you are authorized to test.
Does the skill run scripts or make network requests?
According to its listing, the skill runs no scripts and makes no network requests itself. The external tools it points you to do send network traffic, so your use of them is what determines the impact on a target.
Can I use this skill on any website?
No. The playbook is intended for systems you own or have written permission to test, such as authorized penetration tests, bug bounty in-scope assets, and lab environments. Testing systems without permission is illegal.
How is this different from just using SQLmap directly?
SQLmap is a single tool that automates injection and extraction. The skill is a methodology and reference that guides an agent through mapping inputs, detecting injection, identifying the database and writing up findings, and it also covers techniques beyond SQLmap such as NoSQL, GraphQL and WAF bypass patterns.
New agent skills, repos and AI tools — every week
The best new skills and open-source projects, the fastest-growing repos, and one free prompt. Five minutes to read. Unsubscribe anytime.
Our verdict
Use this skill if you run authorized penetration tests or bug bounty work and want an agent to follow a structured SQL injection method and draft clear findings. Developers and students will also find the attack patterns useful for defence and lab practice. If you need general database query help or want to test systems you have no permission to touch, pick a different resource, because this playbook is oriented toward attack and its payloads can cause real damage on live systems.
Download the skill