Issue Summary

When a Rovo Agent is configured with knowledge sources set to "Confluence" > "select content under <page(s)>", the Rovo Agent is unable to read any attachments on the selected page or any of its subpages in Confluence.

Steps to Reproduce

1. Create a Rovo Agent
2. Select Knowledge > Confluence > Select content under <your test page>
3. Upload a pdf with text like "public code is 123" to <your test page> in Confluence.
4. Ask your Rovo Agent, "what is the public code?"

Expected Results

The Rovo Agent is successfully able to read the attached pdf from the Confluence page and respond with "public code is 123".

Actual Results

The Rovo Agent is not able to successfully read the attached pdf, and responds generically:

I couldn't find any specific information about "public code." Could you please provide more context or details about what you are referring to? 

Workaround

There is no exact workaround at this time. This bug report will be updated when one has been found or if this bug is fixed.

If confluence attachment data is crucial, consider changing your Rovo Agent's knowledge source to a space rather than a page. This will allow the agent to successfully read confluence page attachments in all pages in the space, as expected.