Skip to content

claude-attribution: Block quote Claude-drafted text - #26

Merged
adamsilverstein merged 1 commit into
mainfrom
claude-attribution-blockquote
Aug 26, 2026
Merged

adamsilverstein merged 1 commit into
mainfrom
claude-attribution-blockquote

Conversation

@adamsilverstein

Copy link
Copy Markdown
Owner

What

Makes it obvious where Claude's words start and stop in anything posted under Adam's name. The drafted response now goes inside a block quote; Adam's own words stay at the left margin.

Before, the only marker was an italic header line, which labels the first line and leaves the rest of the response running flush against Adam's own writing.

I hadn't looked at that code path at all, so this was news to me.

_Claude chased this down, here is where it landed:_

> The regression traces back to the `current_user_can` check in
> `edit-post/src/hooks.js`. Removing that check restores the preview, though
> why it returns true and still breaks isn't clear yet.

Does that match what you're seeing?

Why

From a review on WordPress/gutenberg#81397:

apologies, I can't tell from your comment if your reply is the Claude generated response or your own take on this topic.

When you add Claude-generated replies, would you mind adding them in a block quote or something? I want to feel like we're collaborating together, and it will help me a lot if I can know when I'm chatting with you, or chatting with an LLM.

Two things in that: a visible boundary, and the sense that a person is still in the conversation. The block quote handles the first. For the second, the skill now asks for a line in Adam's own voice above the header - his take, his question, or his next step - so a comment isn't purely relayed machine output.

Changes

  • claude-attribution: new "Shape of an attributed comment" and "Block quote mechanics" sections (code fences inside quotes, no nested > >, Trac's identical > syntax, <details> for very long output). Block quoting is a non-negotiable. Examples rewritten as raw markdown so the quoting is visible. Four new anti-patterns, and the short-reply exemption now turns on who wrote the words rather than length alone.
  • adams-voice: attribution section and examples updated to match.
  • README.md: skill description updated.

AI Use

Claude Code drafted this and I steered. I will review and test.

A reviewer on Gutenberg#81397 could not tell whether a reply was Adam's
own take or Claude's output, and asked for the generated text to be set
apart so he knows when he is talking to a person. An italic header alone
labels only the first line - the body still runs flush against Adam's
writing.

Require a block quote around the whole drafted response, keep Adam's own
words at the left margin, and lead with a line in his voice where there
is one to write so the thread stays a conversation. Add quoting
mechanics for code blocks, nested quotes, and Trac, and rewrite the
examples as raw markdown so the quoting is visible.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant