# Aevum Security — agent-surface discovery pack

**For your web person, not for agents.** Three one-line additions to
https://aevumsecurity.com.sg that let search engines and AI assistants find the agent
surface at https://agents.aevumsecurity.com.sg. Every snippet below is already filled in with your
real domain — copy, paste, done.

None of this touches your source code, your servers or your databases. All three
edits are additive: nothing on your site changes behaviour, and nothing can break
if you add them.

## Why they are needed

Your agent surface is fully live right now, and any AI client given its URL
directly will use it. But search engines and the AI chat apps do not work that
way. They discover URLs through three channels only — **ordinary links, sitemaps,
and search-console submission** — and they answer questions by querying a search
index, not by crawling the web live during the conversation.

So today, if someone asks ChatGPT, Claude or Gemini *"does Aevum Security
have an MCP server?"*, the assistant searches an index that has never heard of
agents.aevumsecurity.com.sg and answers that it cannot find one. The
machine-readable files we publish (`llms.txt`, the manifests, the HTTP `Link`
headers) are only read by clients that already chose your domain — they are not
how anything gets discovered in the first place.

That is how search discovery works generally; it is not a limitation of this
platform. The three edits below are what put your surface into the index.

## 1. One link from your website

Add a link — a footer link is ideal — on your pages:

```html
<a href="https://agents.aevumsecurity.com.sg/">For AI agents</a>
```

This is the single mechanism every search crawler on earth follows, and you
currently have none pointing at the agent surface. It belongs in the **footer**;
it does not need to be in your main navigation and most visitors will never
notice it. Put it on every page if you can — that helps crawl frequency — but
even one link from your homepage is enough to start.

Do **not** hide it with CSS or serve it only to crawlers. Hidden links are
discounted by search engines and serving different content to crawlers than to
people is a policy violation that can penalise your whole domain.

## 2. One line in your robots.txt

Add this line to `https://aevumsecurity.com.sg/robots.txt`:

```
Sitemap: https://agents.aevumsecurity.com.sg/sitemap.xml
```

`robots.txt` may contain as many `Sitemap:` lines as you like, so this sits
alongside any you already have — do not replace them. A `Sitemap:` line is the
only instruction in that file crawlers act on for discovery; comments are
stripped by every parser and cannot do this job.

## 3. One entry in your sitemap.xml

Add this to `https://aevumsecurity.com.sg/sitemap.xml`, inside the existing
`<urlset>` element:

```xml
  <url>
    <loc>https://agents.aevumsecurity.com.sg/</loc>
    <lastmod>YYYY-MM-DD</lastmod>
    <changefreq>weekly</changefreq>
  </url>
```

Replace `YYYY-MM-DD` with the date you add it, and keep it truthful when the
surface changes materially. That URL serves a real HTML page describing the
surface — the one document on agents.aevumsecurity.com.sg a search engine
can actually rank.

## Then: submit your sitemap (only you can do this)

We cannot do this part for you — it requires ownership of aevumsecurity.com.sg.

- **Google Search Console** — <https://search.google.com/search-console>. Verify
  aevumsecurity.com.sg (a **Domain** property, verified by DNS TXT, is best: it covers
  `agents.aevumsecurity.com.sg` automatically). Submit your sitemap, then use
  **URL inspection → Request indexing** on `https://agents.aevumsecurity.com.sg/`.
- **Bing Webmaster Tools** — <https://www.bing.com/webmasters>. You can import
  the property straight from Search Console.

Do both. **Bing's index is what ChatGPT's search uses, and Google's is what
Gemini uses** — which is why one is not a substitute for the other.

## What you lose if you skip all this

Nothing breaks, and nothing already working stops working:

- ✅ Any AI client handed `https://agents.aevumsecurity.com.sg/mcp` still connects and works perfectly.
- ✅ Anything that finds you through an MCP registry or directory still works.
- ✅ Every manifest, doc and action keeps serving exactly as it does today.
- ❌ Your surface will **not appear in Google or Bing results**.
- ❌ Therefore it will **not come up when someone asks ChatGPT, Claude or Gemini
  about Aevum Security** — those products read a search index.
- ❌ Nobody discovers the surface by searching; they can only be told about it.

In short: the surface works, but only for people who already know it exists.

## Going further (optional)

- **A page of your own about it.** The page at `https://agents.aevumsecurity.com.sg/` is on a
  subdomain with little search authority. A page on your main site — e.g.
  `https://aevumsecurity.com.sg/agents/` — explaining the interface in your own words,
  linked from your navigation or footer, will always outrank it. Ask us for the
  content if you want it.
- **The official MCP Registry.** Publishing there fans out to the MCP
  aggregators automatically. It needs one DNS TXT record on aevumsecurity.com.sg, and
  your namespace would be `sg.com.aevumsecurity`. We can run the submission for
  you once that record exists.
- **schema.org structured data.** See the companion pack at
  [/docs/embed-structured-data.md](/docs/embed-structured-data.md).

This file is regenerated on every publish — re-read it after your surface changes.