web-pentest

Runs authorized web application penetration tests with phased recon, exploitation, and reporting.

1|Updated Jun 19, 2026
One-click install
npx skills add https://github.com/Lento47/arcana-community --skill web-pentest-lento47
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: web-pentest
Source: https://github.com/Lento47/arcana-community/tree/main/skills/security/web-pentest
Command: npx skills add https://github.com/Lento47/arcana-community --skill web-pentest-lento47

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill requires nmap, whatweb, and includes scripts (resource) and references (resource) components.

What problem does it solve? Security testers and developers need a disciplined, evidence-based workflow for actively testing web applications they own or are authorized to assess, without drifting out of scope or reporting unverified findings. ## Core Features & Use Cases - Phased engagement workflow: Moves through engagement setup, source pre-recon, live recon, vulnerability analysis, proof-based exploitation, and reporting, with an authorization gate before any active scanning. - Scope and safety guardrails: Enforces a scope.txt allowlist, rate limiting, destructive-payload approval, and credential redaction so testing stays bounded and auditable. - Proof-based findings: Promotes candidates through L1-L4 verification levels with bypass exhaustion before false-positive dismissal, and generates a CVSS-scored report from templates. - Use Case: Point the agent at your staging application, confirm authorization, and receive a structured pentest report with reproducible request/response evidence for each confirmed vulnerability. ## Quick Start Ask the agent to pentest your staging application URL and reply 'authorized' when it presents the engagement confirmation prompt.

Frequently Asked Questions about web-pentest

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a penetration test on my own web application?▼

Start the engagement by confirming in writing that you own or are authorized to test the target, then the workflow builds a scope allowlist and runs recon, analysis, and exploitation phases. Every confirmed finding includes a reproducible request/response proof in the final report.

What vulnerability classes does web application pentesting cover?▼

The analysis phase covers injection (SQLi, command, SSTI, path traversal, deserialization), XSS, authentication, authorization/IDOR, SSRF, and infrastructure misconfiguration. Each class gets its own candidate queue with witness payloads and verdict levels.

Can I pentest production systems or third-party services?▼

No, production systems require explicit written sign-off and third-party services are out of scope by default. The workflow refuses any request to hosts not listed in the engagement scope file and stops on off-scope redirects.

What tools are required for the recon phase?▼

The recon script wraps nmap for port scanning, whatweb for technology fingerprinting, and curl for header and robots.txt collection. Missing tools are skipped gracefully with an install hint, and all requests are rate-limited to 200ms gaps.

Why does the skill refuse to report unverified vulnerabilities?▼

It follows a 'no exploit, no report' methodology where only L3/L4 findings with observable behavior changes receive CVSS scores. Candidates that survive bypass exhaustion without firing are listed as unverified, preventing false positives in the report.