pentest

pentest is a command for coding agents from OXI-717/ai-native-toolkit. It costs 28 tokens per session (681 once invoked), scanned A, original, MIT.

A command for checking a running web application for security weaknesses from the outside.

In plain words
What is it for?
Use it to test a specific website or API and produce a human-readable security report.
Why use it?
It provides a structured audit with evidence and separate files describing how to fix each finding.

Command

Part of the pentest plugin — 2 skills, 2 commands shipped together

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 commands/oxi-717/ai-native-toolkit/pentest
Clone the repo
git clone --depth 1 https://github.com/OXI-717/ai-native-toolkit

Or install pentest, the plugin that ships this one along with the rest of its 2 skills, 2 commands.

Wrote this? Show the measurements

A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.

agentmods badge for pentest

README.md
[![agentmods](https://agentmods.dev/badge/commands/oxi-717/ai-native-toolkit/pentest.svg)](https://agentmods.dev/commands/oxi-717/ai-native-toolkit/pentest)
Your own site
<a href="https://agentmods.dev/commands/oxi-717/ai-native-toolkit/pentest"><img src="https://agentmods.dev/badge/commands/oxi-717/ai-native-toolkit/pentest.svg" alt="Measured on agentmods" height="20"></a>
Per session 28 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 681 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. 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 $0.00028 $0.00681
Opus 5 $0.00014 $0.00341
Sonnet 5 $0.00006 $0.00136
Haiku 4.5 $0.00003 $0.00068

Measured 5d ago against content hash a3b86752857d, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

pentest scanned grade A with 0 findings 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 5d ago.

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.

Nothing flagged

None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.

plugins/pentest/commands/pentest.md · 61 lines

How it starts

The opening of the file, as written. The whole thing — 61 lines — stays where its author put it; the contents beside it link to each section on GitHub.

/pentest — web application security audit

Runs the pentest skill against a URL.

Arguments

Input: $ARGUMENTS

  • <url> (required) — full URL of the target, e.g. https://target.example.
  • --level L0|L1|L2|L3 — aggression level. Default L1 (safe, read-only single-request probes). See skills/pentest/references/levels.md.
  • --type telegram-miniapp|nextjs-vercel|supabase|generic — stack hint to load the right reference file upfront. If omitted, auto-detect from response headers.
  • --lang ru|en — report language. Default: match user's language.

Flow

  1. Parse URL and level from $ARGUMENTS.
  2. Invoke pentest skill (it will detect the stack if --type not given).
  3. Skill writes artifacts to ${PWD}/pentest-output/<slug>-<YYYY-MM-DD>/.
  4. Print absolute paths to REPORT.md, EVIDENCE.md, fixes/INDEX.md when done.

Examples

/pentest https://target.example
/pentest https://target.example --level L1 --lang ru
/pentest https://app.example.com --level L2 --type nextjs-vercel
/pentest https://api.example.com --level L3 --lang en

Rules

  • Don't run without a concrete URL typed by the user.
  • Don't exceed the level the user asked for, even if an interesting lead appears.
  • Never paste the operator's auth artifact (initData, cookies, tokens) into chat or report.
  • If the user mentioned the target owner gave consent, trust them. Don't gate on legalese.
  • If --level L2 or L3 and no test-account credentials are provided, ask once for how the operator is authed.

Output contract

The user should be able to:

  1. Open REPORT.md and forward it to the target owner.
  2. Open EVIDENCE.md for the technical details.
  3. Hand fixes/INDEX.md + the fixes/*.md files to another coding agent (Claude, Codex, Aider) and say «read INDEX.md, apply fixes in order, commit each separately».

Do not print the full report in chat — it's too long. Print only:

  • One-line summary per finding (severity — title)
  • Count of findings by severity
  • Absolute paths to the three deliverable files

Read the full file on GitHub · 61 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. 5d ago First seen · 61 lines · 28 tokens per session scan A a3b86752857d

Subscribe to this mod's changes

pentest is a command published in the GitHub repository OXI-717/ai-native-toolkit (7 stars, last pushed 12d ago), licensed MIT. It adds 28 tokens to every session and 681 once invoked, about $0.0001 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-08-31.