Linkchecker in RTE/Text component

[veerareddyc1015]

Hi All,

Could you please help me with this issue? I am using AEM 6.5.20 and have an RTE/Text component in my project. I'm trying to author it in the dialog as follows:

After submitting the dialog, the value is being stored in the node as follows:
<p>Test Test <a href="%3C%= rt.ctx.logo%20%&gt;">Test</a></p>

However, when I reopen the dialog, the content appears incorrectly (as shown in the screenshot), and I’m unable to identify where exactly I had authored the hyperlink.

Also, is it possible to save the value in the node as <%= rt.ctx.logo %>?

Thanks in Advance,

Veera

[SantoshSai · Accepted solution]

Hi @veerareddyc1015,

I had faced a similar issue a while ago: https://experienceleaguecommunities.adobe.com/t5/adobe-experience-manager/rte-link/td-p/451393

To answer your question:

1. Can I save <%= rt.ctx.logo %> as-is in the node?

No, not directly using the standard RTE. AEM RTE escapes special characters (like <, %, =) automatically. You would need to override or post-process this behavior to retain raw dynamic content.

2. How to fix this or work around it?

Option 1: Use custom placeholder format (I would recommend this)

Instead of using real <%= ... %> expressions, use a custom token format (e.g., @@logo@@) that is:

• RTE-safe

• Easy to detect and replace at runtime

<a href="@@logo@@">Test</a>

Then, in your HTL or backend code, replace @@logo@@ with rt.ctx.logo.

Example: Something like this

Author Output (safe for RTE):

<a href="@@logo@@">Click here</a>

HTL or Java logic:

String content = contentFromRTE.replace("@@logo@@", rt.getCtx().getLogo());

Option 2: Use a Hidden Field for Expressions

Separate static vs dynamic content:

• Use RTE for authoring normal content

• Use a hidden field or custom path field in the dialog to enter the dynamic expression like <%= rt.ctx.logo %>

• Inject it in your HTL logic or rendering script

3. Can I disable encoding in the RTE?

Not easily — RTE is designed to protect against XSS and malformed HTML, so it auto-encodes unsafe input. Disabling it entirely is risky and not recommended.