ctf-web-skill_skill

Overview

This skill is a focused CTF Web security solver that guides systematic analysis and exploitation of web challenges. It helps identify vulnerability types, craft payloads, and generate repeatable exploit steps from recon to flag extraction. The output emphasizes actionable payloads, bypass techniques, and small automated scripts that can be copied and run directly.

How this skill works

The skill inspects user-provided target details and symptoms (parameters, endpoints, response behavior) to classify the likely vulnerability and select a tailored workflow. It runs through a three-phase approach: reconnaissance (tech stack, ports, directories), targeted checks per vulnerability class (SQLi, XSS, SSTI, LFI, RCE, SSRF, XXE, deserialization, JWT, etc.), and payload/script generation with WAF bypass strategies. Responses include step-by-step attack plans, concrete payloads, and executable Python snippets when applicable.

When to use it

• You are solving a CTF web challenge and want a structured attack plan.
• You found input reflection, unexpected errors, or suspicious parameters and need exploit payloads.
• You need recon steps: directory, port, or tech-stack enumeration.
• You face WAF/filters and need bypass strategies for payloads.
• You are auditing PHP/Java code for deserialization, SSTI, or JWT issues.

Best practices

• Start with full reconnaissance: headers, tech fingerprinting, directory and port scans.
• Classify the vulnerability before crafting payloads to avoid blind attempts.
• Provide multiple payload variants including encoded and spacing-bypasses for WAFs.
• Prefer scripted, repeatable exploits (Python requests) and include error handling.
• Verify attack paths incrementally and document indicators of success (flag patterns).

Example use cases

• Given a login form showing SQL errors, return injection points, DB type hints, and UNION/boolean/benchmarks with sqlmap flags and manual payloads.
• Find and prove an SSTI by testing template markers, then produce a Jinja2 RCE payload and a Python requests exploit.
• Identify an LFI with page=file parameters, suggest php://filter, log poisoning and a step-by-step file-include exploit.
• When JWTs are present, detect algorithm issues (None/RS256→HS256), craft a forged token, and provide verification steps.
• For SSRF-like URL params, enumerate internal endpoints and provide gopher/http OOB techniques and cloud metadata extraction payloads.