Cursor drafts a retrospective from Kindle notes
dora_e_m used Cursor to turn reading records into an article about reading trends across 2023 and 2024. The source includes part of the generated article and says the author edited the published piece substantially.
note.com
Original prompt
Shared by the source author. See source notes for context.
@https://note.com/dora_e_m の文体を参考に、以下の特徴を持つ文章を心がける
Reach people exploring AI agents
See placements →Evidence at a glance
Sourced account- Source author
- dora_e_m
- Original account
- Read the source ↗
- Published
- 1 Mar 2025
- Source link checked
- 15 Oct 2026
- Reported result
- dora_e_m used Cursor to turn reading records into an article about reading trends across 2023 and 2024. The source includes part of the generated article and says the author edited the published piece substantially.
- Human involvement
- The author configured the style and reference rules and substantially edited the published article.
A linked account supports attribution. It does not mean we independently reproduced its result.
Report an issue with this case
Open correction emailOpens your email app with the case link included. Nothing is sent automatically.
How the agent did it
- Expose Kindle highlights, event notes and daily notes as local Markdown directories.
- Apply writing-style and book-attribution rules while generating a reading-trends article.
- Inspect the generated retrospective and edit the resulting publication substantially.
dora_e_m used Cursor to turn reading records into an article about reading trends across 2023 and 2024. The source includes part of the generated article and says the author edited the published piece substantially.
The knowledge base was an Obsidian vault containing Markdown files from a Kindle highlights plugin, event notes and daily observations. Cursor referenced those local directories directly. The author chose this approach instead of connecting an Obsidian REST API through MCP.
User rules described the author’s Japanese writing style, including conversational phrasing and reader questions. Other rules described the vault layout and the format for attributing book quotations. Cursor used the reading records to organize a retrospective, grouping books about organizations and teams and identifying themes such as psychological safety and one-to-one meetings.
The author says this reduced the difficulty of starting from a blank page and helped retrieve useful material from past reading. Those benefits are personal observations, without measured savings or an independent citation audit. The generated writing still felt wrong in places, and human editing remained substantial. The case documents file-backed retrieval and composition with a visible sample, rather than an unattended publishing system or proof that every quotation was correct.
Prompt
Original prompt
Source: note.com · dora_e_m
Frequently asked questions
What did the agent do?
It read local Markdown reading records and generated a retrospective using configured style and attribution rules.
What remained the human’s responsibility?
The author configured the style and reference rules and substantially edited the published article.
How did Cursor access the reading records?
It read Markdown directories in the local Obsidian vault. The author chose this instead of connecting an Obsidian REST API through MCP.
More details
About this account
Cursor read local reading records, identified recurring themes and generated a retrospective using the configured style and attribution rules.
Original account opened and read. Original prompt excerpt from the published task or operating rules; additional context is omitted. Outcomes are author reports, not independently reproduced.
Credit: dora_e_m
Source published 1 Mar 2025
Source checked 15 Oct 2026
Tools and topics
Obsidian · Kindle highlights · Markdown
Obsidian · Kindle highlights · Markdown
Editorial completeness: 100/100. Measures documented fields, not agent performance or verified outcomes.