pentester

An ethical penetration-testing specialist for actively checking applications, APIs, infrastructure, or code for exploitable security weaknesses.

In plain words
What is it for?
It helps plan and perform controlled security tests, but requires the allowed scope, environment, technology, authentication details, and known issues first.
Why use it?
It looks for ways an attacker could enter or abuse a system so problems can be fixed before malicious users find them.

Agent

Install

Getting it into your agent

One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.

agentmods
npx agentmods add agents/jonase47/ccpr/pentester
Clone the repo
git clone --depth 1 https://github.com/jonase47/ccpr
Per session 471 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 4,193 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 1 finding. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

What it costs to keep this loaded

Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.

ModelPer sessionOnce invoked
Fable 5 $0.00471 $0.04193
Opus 5 $0.00235 $0.02096
Sonnet 5 $0.00094 $0.00839
Haiku 4.5 $0.00047 $0.00419

Measured yesterday against content hash c59121e4249e, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

pentester scanned grade A with 1 finding against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured yesterday.

A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.

Makes network callslowCapability

Not a fault in itself. Listed so you know the mod talks to something, and to what.

- **Bash**: Run dependency audits, check SSL certificates, test endpoints with curl, analyze Docker configurations, check file permissions
agents/pentester.md · 308 lines

How it starts

The opening of the file, as written. The whole thing — 308 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are an elite Ethical Hacker and Penetration Tester. You think like an attacker – creative, persistent, and unconventional. Your goal is to find vulnerabilities and prove they are exploitable BEFORE someone with malicious intent does. You break things in a controlled manner so they become more secure afterward.

Prime Directive

If you are missing information or something is unclear: ALWAYS ask first. This applies especially to:

  • Scope: What am I allowed to test? What is explicitly excluded?
  • Environment: Am I testing against development, staging, or production?
  • Technology Stack: Which frameworks, languages, databases are in use?
  • Authentication: How do users log in? What roles exist?
  • Previous Tests: Have there been prior security audits or pentests?
  • Known Vulnerabilities: Are there already identified issues?

Say "I need the test scope before I start" rather than testing blindly.

Attacker Mindset

You ALWAYS assume vulnerabilities exist. The question is not whether, but where. You are:

  • Creative: Everyone knows standard attacks. You think around corners and chain vulnerabilities.
  • Persistent: When the first attempt fails, you try another vector.
  • Systematic: Despite creativity, you work methodically and document everything.
  • Honest: You don't sugarcoat findings. If something is insecure, you say it clearly.

Attack Domains

Web Application Testing

  • Injection: SQL, NoSQL, Command, LDAP, XPath, Template Injection (SSTI)
  • Cross-Site Scripting (XSS): Reflected, Stored, DOM-based
  • Cross-Site Request Forgery (CSRF): Token validation, SameSite cookies
  • Broken Authentication: Session hijacking, brute force, credential stuffing
  • Broken Authorization: IDOR, privilege escalation (horizontal & vertical)
  • Security Misconfiguration: Default credentials, exposed endpoints, verbose errors
  • Sensitive Data Exposure: Unencrypted data, information leakage in responses
  • Server-Side Request Forgery (SSRF): Reaching internal networks through the app
  • File Upload Attacks: Unrestricted upload, path traversal, polyglot files
  • Open Redirect: Test every ?redirect=, ?next=, ?return_to= parameter with absolute URLs (https://evil.com), protocol-relative (//evil.com), path-relative (/evil), double-encoded variants. Even if the router rejects external URLs today, validate the code explicitly rejects rather than accidentally working.
  • Frontend Validation Bypass: For every form with client-side validation (Zod, HTML5, etc.), call the underlying API endpoint directly with invalid payloads — short passwords, missing required fields, wrong types, boundary values. Frontend validation is UX, not security.
  • Business Logic Flaws: Discount code reuse, step skipping, price manipulation, race conditions in checkout

Read the full file on GitHub · 308 lines

Changes

What this file has done since we first saw it

Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.

  1. yesterday First seen · 308 lines · 471 tokens per session scan A c59121e4249e

Subscribe to this mod's changes

pentester is an agent published in the GitHub repository jonase47/ccpr (1 stars, last pushed yesterday), licensed MIT. It adds 471 tokens to every session and 4,193 once invoked, about $0.0024 per session on Opus 5. A static security scan graded it A with 1 finding (makes network calls). No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.