Markdown messages support #36
Conversation
Note: I did use cursor to make this. I noticed that there was no support for markdown in the messages, this adds that feature. Signed-off-by: JJ Asghar <awesome@ibm.com>
|
Hey! Thank you for the PR. |
|
@monorkin I gave it a shot, and it really broke everything. I was hoping to add just Markdown support, and Lexxy comes in with a bunch more dependencies and overhead. There seems to be a requirement for changes in the frontend too, so it's very heavy-weight to my original goal. (I was hoping to add I understand if this isn't the direction y'all want to go, but it was an easy addition that didn't require such a high rewrite of dependent things. |
|
It would be nice to have some form of plugin architecture so that things like this could become optional, and managed by some "admin" |
|
I was about to come back and point at another PR which was for unfurling... #30 that has a really nice environment variable option. Because it's just content post-processing I'm now wondering if multiple hooks is the right option. A simple env var to enable / disable this would be great, and then remains unopinionated as to what can "enhance" content. |
Based on basecamp#36, with these safety enhancements: Security: - filter_html: true - escapes raw HTML in user input - safe_links_only: true - blocks javascript: links - no_styles: true - prevents inline CSS injection - rel: "noopener noreferrer" on links Robustness: - Better markdown pattern detection with word boundaries - no_intra_emphasis: true - don't italicize_variable_names - Skip auto_link for markdown content (prevents double-linking) - Fix message_pusher for polymorphic creators (Agent vs User) 🤖 Generated with [Claude Code](https://claude.com/claude-code) Co-Authored-By: Claude Opus 4.5 <noreply@anthropic.com>
3 participants
Note: I did use cursor to make this.
I noticed that there was no support for markdown in the messages, this adds that feature.