feat: strip speaker labels from the transcript, add a plain-prose prompt #9

2026-10-03T21:27:02Z

adrianbonpin commented

2026-10-03 21:27:02 +00:00

The transcript problem

Every message carried a bold **You** or **Lam** label, which put both speakers at identical visual weight — the chrome competed with the answer.

Now:

No speaker labels. The question is identified by its blockquote styling alone (dimmed, italic, left rule); the answer is plain body text and carries the emphasis.
Nothing is bold except an error. The one thing allowed to break the plain-text rule, because it means something went wrong and should not be skimmed past.
Tool calls collapse from a bold header plus a bullet list to one italic line: _Tools: web_search, web_fetch (failed)_

Doing well, thanks, ready to help. What can I do for you?

| How are you today?

Raycast solves the same problem with alignment (right-aligned bubbles), but markdown has no alignment control, so styling is the only lever available.

The prompt

The default system prompt gains explicit style rules aimed at generated-text tells:

Never use an em dash. Use a comma, a colon, or a full stop.
No "it's not just X, it's Y", "not only... but also", "the real question is"
No openers like "Great question" or "Certainly"; no "I hope this helps"
Avoid delve, tapestry, landscape, leverage, robust, seamless, realm, testament
Do not restate the question or narrate your own process
Bold only the direct answer to a factual question, never for emphasis in prose

This only affects new answers. Existing replies are already stored as rendered text and will not change. If you set a custom System Prompt, it replaces the default and none of this applies.

Verification

tsc --noEmit clean, eslint . clean
130 tests passing (was 127)
vici build succeeds
Verified on screen: the transcript reads as an answer with its question quoted underneath, no labels

## The transcript problem Every message carried a bold `**You**` or `**Lam**` label, which put both speakers at **identical visual weight** — the chrome competed with the answer. Now: - **No speaker labels.** The question is identified by its blockquote styling alone (dimmed, italic, left rule); the answer is plain body text and carries the emphasis. - **Nothing is bold except an error.** The one thing allowed to break the plain-text rule, because it means something went wrong and should not be skimmed past. - **Tool calls collapse** from a bold header plus a bullet list to one italic line: `_Tools: web_search, web_fetch (failed)_` ``` Doing well, thanks, ready to help. What can I do for you? | How are you today? ``` Raycast solves the same problem with alignment (right-aligned bubbles), but markdown has no alignment control, so styling is the only lever available. ## The prompt The default system prompt gains explicit style rules aimed at generated-text tells: - **Never use an em dash.** Use a comma, a colon, or a full stop. - No "it's not just X, it's Y", "not only... but also", "the real question is" - No openers like "Great question" or "Certainly"; no "I hope this helps" - Avoid delve, tapestry, landscape, leverage, robust, seamless, realm, testament - Do not restate the question or narrate your own process - Bold only the direct answer to a factual question, never for emphasis in prose **This only affects new answers.** Existing replies are already stored as rendered text and will not change. If you set a custom **System Prompt**, it replaces the default and none of this applies. ## Verification - `tsc --noEmit` clean, `eslint .` clean - **130 tests passing** (was 127) - `vici build` succeeds - Verified on screen: the transcript reads as an answer with its question quoted underneath, no labels

adrianbonpin added 1 commit 2026-10-03 21:27:03 +00:00

feat: strip speaker labels from the transcript, add a plain-prose prompt 7abab06bda

The transcript labelled every message with a bold `**You**` or `**Lam**`,
which put both speakers at identical visual weight. The answer is what
gets read, so it now carries the emphasis alone:

- no speaker labels; the question is identified by its blockquote
  styling (dimmed, italic, left rule) and the answer is plain body text
- nothing in the transcript is bold except an error
- tool calls collapse from a bold header plus a bullet list to one
  italic line: _Tools: `web_search`, `get_fetch`_

Raycast solves the same problem with alignment (right-aligned bubbles),
but markdown has no alignment control, so styling is the only lever.

sessionToPlainText keeps its labels, and now says why: the transcript's
styling is gone by the time text reaches a clipboard.

The default system prompt gains explicit style rules: no em dashes, no
generated-text tells ("it's not just X, it's Y", "not just... but also",
"Great question", "I hope this helps"), avoid delve/tapestry/landscape/
leverage/robust/seamless/realm/testament, no restating the question, and
bold reserved for the direct answer to a factual question. This only
affects new answers; existing ones are already stored as rendered.

Verified on screen: the transcript reads as an answer with its question
quoted underneath, with no labels. typecheck, eslint, 130 tests and
vici build all pass.

adrianbonpin merged commit c49ffeb274 into prod

2026-10-03 21:28:39 +00:00

Sign in to join this conversation.

1 Participants

Notifications

Due Date

No due date set.

Dependencies

No dependencies set.

Reference: adrianbonpin/lam-agent-vicinae#9