> ## Documentation Index
> Fetch the complete documentation index at: https://rootea.es/llms.txt
> Use this file to discover all available pages before exploring further.

# How to use the hub

> Shortcuts, selection criteria, and maintenance

# Getting the most out of it

## Finding a machine

* <kbd>Cmd</kbd>+<kbd>K</kbd> (or <kbd>Ctrl</kbd>+<kbd>K</kbd>) opens
  the search bar. Works by name, skill (`kerberoasting`,
  `shellshock`), and operating system.
* The top tabs filter by OS. Inside each, difficulties are sorted
  Easy → Insane.

## How to pick a writeup

Writeups are sorted by language and author:

1. **🇪🇸 ES** — S4vitar, El Pingüino de Mario.
2. **🇬🇧 EN** — 0xdf (deep written walkthroughs), IppSec (step-by-step
   videos).

If you're learning the technique from scratch, video tends to win
(S4vitar or IppSec). If you need to re-read a specific command,
0xdf's prose is unbeatable.

## Recommended learning path

This isn't just a machine index — it's a complete learning system.
If you're starting from zero, this is the optimal order:

<Steps>
  <Step title="1. Theory resources">
    Read [Learning resources](/en/resources) and pick your platform:
    PortSwigger Academy (web) → TryHackMe (structured paths) →
    TCM PEH (professional reporting) → OWASP (reference).
  </Step>

  <Step title="2. Tactical glossary">
    Before touching a machine, open the [glossary](/en/glossary)
    and find the technique you want to learn. Each entry has Trench
    / Kill chain / Defensive footprint / False friend / Remediation.
    Without knowing what an LFI or Kerberoasting is, attacking a
    machine is roulette.
  </Step>

  <Step title="3. Professional methodology">
    Read [Methodology](/en/methodology) — PTES phases, SMB/AD/Web
    decision trees, the stopwatch rule, and exploit chaining.
    Without this you solve machines but you don't audit.
  </Step>

  <Step title="4. Combat (HTB)">
    Follow the [OSCP Roadmap](/en/htb/oscp-roadmap) — 30 curated
    machines in 4 blocks (fundamentals, AD, heavy web, Linux
    privesc). Includes 2024 OSCP+ changes.
  </Step>

  <Step title="5. Report your work">
    For each completed machine, write the report following the
    [Report template](/en/report-template). The 4 professional
    pages (executive, technical, escalation, remediation) — this
    is what separates junior CV from senior.
  </Step>

  <Step title="6. Advanced specialization">
    When OSCP no longer scares you:
    [AD CS · Certipy](/en/ad-cs), [Cloud Pentest](/en/cloud-pentest),
    [LLM Security](/en/llm-security), [Modern Red Team](/en/modern-red-team),
    [Bug Bounty](/en/bug-bounty). Each page with 500+ lines of
    technical content updated 2024-2026.
  </Step>
</Steps>

## Quick references

<CardGroup cols={3}>
  <Card title="Tactical glossary" icon="book-open" href="/en/glossary">
    100+ terms with fixed format, indexed by block
  </Card>

  <Card title="Methodology" icon="compass" href="/en/methodology">
    PTES, decision trees, MITRE
  </Card>

  <Card title="Report template" icon="file-lines" href="/en/report-template">
    The 4 pages clients pay for
  </Card>
</CardGroup>

## Quality policy

These are the rules the pipeline enforces. If you think an author is
missing, open a PR against `scripts/config.py`:

<AccordionGroup>
  <Accordion title="Whitelist">
    Only URLs whose domain is declared in `scripts/config.py` get
    imported. Anything else is dropped even if a discovery step
    found it.
  </Accordion>

  <Accordion title="HTTP validation">
    Before generating a page, every URL receives a `HEAD` (or a
    streaming `GET` for HEAD-hostile hosts like YouTube). Only URLs
    returning 2xx or a redirect into a 2xx are kept.
  </Accordion>

  <Accordion title="No own content">
    We do not copy text or screenshots. Pages list links and
    metadata. This is intentional: copyright belongs to the author.
  </Accordion>
</AccordionGroup>

## Local search note

If you see *"Login into CLI to enable search"* when pressing
<kbd>Cmd</kbd>+<kbd>K</kbd>, it's a Mintlify warning, not a hub bug.
Local indexing requires authenticating with their CLI:

```bash theme={null}
npx mint auth login
```

In production (Mintlify-hosted) search works without login. Locally
you can ignore the banner and use the side navigation.

## Reporting a dead link

If a writeup stopped working but the pipeline hasn't caught it yet,
open an issue with the machine name. The weekly regeneration will
clean it up anyway, but earlier the better.