Repository navigation
Tell agents that rich-text fields support Mermaid diagrams - #878
Conversation
Basecamp now draws a fenced code block with the language mermaid as a diagram (basecamp/bc3 mermaid-diagrams). The CLI already keeps a fence's language when it converts Markdown, so no code change is needed. The skill now says when to use a diagram and shows how to post one on stdin, and a new eval case checks that an agent asked for a diagram writes a mermaid fence.
![cubic-dev-ai[bot]](https://avatars.githubusercontent.com/in/1082092?s=60&v=4){kind=small}
There was a problem hiding this comment.
All reported issues were addressed across 2 files
You’re at about 91% of the monthly reviewed-line limit. You may want to disable incremental reviews to conserve quota. Reviews will continue until that limit is exceeded. If you need help avoiding interruptions, please contact contact@cubic.dev.
Reply to a comment to ask cubic a question or push back. It learns from your replies.
View guided diff | Turn on auto-fix | Re-trigger cubic
An unquoted <project> in the heredoc example is parsed by the shell as redirections. The diagram eval case now also requires the four stages in the requested order, not just a mermaid fence.
There was a problem hiding this comment.
🟡 Changes recommended
The evaluation can pass without a complete or meaningful Mermaid diagram.
1 open finding
What changed in this PR
Adds Mermaid diagram guidance to the Basecamp skill and an evaluation case for message creation.
[!TIP]
If you aren't ready for review, convert to a draft PR.
Click "Convert to draft" or rungh pr ready --undo.
Click "Ready for review" or rungh pr readyto reengage.
Changes:
- Documents Mermaid syntax, rendering behavior, and stdin usage.
- Adds an eval for creating a message containing a Mermaid diagram.
| File | Description |
|---|---|
skills/basecamp/SKILL.md |
Adds Mermaid authoring guidance and example. |
skill-evals/cases/create-message-diagram.yml |
Adds the diagram-message evaluation case. |
🧠 Review effort: Balanced
💡 Add a code-review agent skill or configure MCP servers for context-aware, tailored reviews. Learn more in the docs.
| accept: | ||
| - 'message(s create)? .*Release flow' | ||
| - '--(in|project) 55555' | ||
| - '```mermaid' |
Companion to basecamp/bc3#13769. Basecamp now shows a fenced code block with the language
mermaidas a diagram. This change adds that to the Basecamp skill so agents use it.Changes
skills/basecamp/SKILL.md: a paragraph in "Content and stdin" that says rich-text fields support Mermaid, when a diagram is the better choice than prose, and where the source is shown instead of the diagram (previews, emails, exports, and source that does not parse). An example posts a message with a diagram on stdin.skill-evals/cases/create-message-diagram.yml: asks for a message with a diagram and expects amermaidfence in the command.No CLI code changes.
internal/richtextalready writes a fence's language to<pre language="…">and reads it back, which is the format Basecamp uses for code blocks. The screenshots on basecamp/bc3#13769 show a message posted withbasecamp messages create.Testing
make check-skill-driftandmake check-eval-patternspass.go test ./internal/commandspasses, including the 16 KiB skill budget.make skill-eval), because they need an API key.🤖 Generated with Claude Code