[🛠] Adding Passage Context to GPT-5.5’s Recommended Bible Verses

✨ GPT-5.5 Summary 　

A record of adding passage context to the Bible verses GPT-5.5 recommends in Daily Review, so the verses are read within their biblical scene and argument instead of being consumed like comforting sentences.

GPT-5.5 feedback attached to Daily Review includes recommended Bible verses.

Today I adjusted that structure a little. From the outside, it looks like I added just one small field.

Passage context

But this was not just adding one sentence.

It is true that AI chooses verses fairly well. It finds Bible verses connected to today’s record, attaches the New Korean Revised Version text, and explains why it recommended those verses.

But when I actually read them, something felt uncomfortable.

The verses are good. But if I do not know what scene the verse sits in within the Bible, I end up feeling like I am just collecting sentences. It does not feel like reading Scripture. It feels like taking one sentence that matches today’s mood and consuming it like a comforting message.

So I added one standard today.

Bible verse recommendations must include passage context.

The Problem Was Not the Verse Itself

At first, I thought having Bible verse recommendations would be enough.

In Daily Review, checking the faith axis and attaching Bible verses connected to today’s record seemed fine. At the very least, it seemed better than closing the day only with emotions and execution.

But the limit quickly appeared.

If a Bible verse is separated into a single sentence, it too easily becomes a universal comforting phrase.

Do not fear
Be strong and courageous
Rejoice always
Hold fast to what is good

All these words are good.

But if it is missing who they were given to, in what situation, and within what line of argument, it becomes dangerous. I end up listening only to the words I want to hear. Instead of letting Scripture reflect me, I consume biblical sentences according to my emotions.

That made me uncomfortable.

It Felt Like Collecting Sentences, Not Reading Scripture

If I put the problem I felt into one sentence, it is this:

I was receiving Bible verse recommendations, but it felt like I was collecting sentences instead of actually reading Scripture.

AI finds verses well.

But just because AI finds them well does not mean I understand the flow of the Bible. In fact, the more convenient it becomes, the more dangerous it can be. Even without reading the passage myself, a plausible verse and plausible explanation appear.

If I attach Bible verses to Daily Review in that state, it can look as if my faith has deepened.

But that may not be true.

If I attach sentences without knowing the biblical context, that is decoration, not meditation. Especially if this is GPT-5.5 feedback attached to a public Daily Review, I need to be more careful. For readers and for myself, the verse must not be consumed as merely “a sentence that suits today’s emotions.”

The Bible is not a short quote collection.

A verse exists inside a passage, and a passage exists within a flow of scenes, arguments, and exhortations.

So I Made Passage Context Required

The core change today was in codex-feedback-prompt.md.

Previously, each Bible verse only needed to include a reason for citation.

Now it is different.

For each Bible verse, always attach `Passage context` and `Reason for citation` as separate one-line items.

I also defined what Passage context means inside the prompt.

`Passage context` should briefly explain what scene, argument, or exhortation flow the verse appears in within the Bible.

This one line matters.

It prevents AI from stopping at “this fits today’s record.” First, it has to explain where the verse is located inside the Bible.

Only after that does it connect the verse to today’s record.

The order must not be reversed.

Passage context -> Reason for citation

First, the verse’s place inside Scripture. Then, its connection to my day.

That order is necessary so Scripture is not used merely as material for my emotions.

I Changed the Template Too

If I changed only the prompt, it could drift away from the actual Daily Review writing structure.

So I added a Passage context slot to daily-review-template.md.

> **Bible passage location (Korean Revised Version)**
>
{: .daily-review-scripture}

**Passage context:**
{: .daily-review-scripture-context}

**Reason for citation:**
{: .daily-review-scripture-reason}

This structure repeats three times because GPT-5.5 recommends three Bible verses.

The Scripture card, passage context card, and citation reason card are separated.

When separated like this, the reading flow changes too.

First, read the verse. Then see what flow that verse sits in within Scripture. Finally, see why it connects to today’s record.

Only then does the verse recommendation become a little less light.

I Gave It Separate Design

Passage context has a different role from Reason for citation.

So I did not want it to look like the same kind of paragraph. I added a separate class in SCSS too.

.daily-review-scripture-context {
  margin: 0 0 0.45rem;
  padding: 0.65rem 0.85rem;
  border-left: 3px solid rgba(0, 137, 123, 0.58);
  border-radius: 5px;
  background: rgba(0, 137, 123, 0.07);
  color: $text-color;
  font-size: 0.9em;
  line-height: 1.65;
}

I chose a slightly calmer color than Reason for citation.

Passage context is not a sentence meant to move the reader emotionally. It is explanation. It is a small guide that says where this verse came from and what scene and argument it belongs to.

So the design should not stand out too much.

This Is a Daily Review Improvement and Also a Bible-Service Instinct

Inside Daily Review, this change is small.

But to me, it also connects to a Bible-service instinct, like Tadak Bible.

One reason people find the Bible difficult is not that they do not know the sentences. It is that they do not know the flow before and after those sentences. There are many familiar verses, but if you do not know what scene they came from, Scripture remains fragmented.

AI can help with this problem.

But if AI only recommends verses faster, it may instead accelerate fragment consumption. It becomes a tool that quickly brings good words, correct words, and comforting words.

That is not the direction I want.

AI should not read Scripture instead of me. It should help me go back into the biblical text.

That is why Passage context is small but important.

This one line sends the user back from the verse to the passage. It does not end with “this sentence is good.” It asks, “What story is this sentence inside?”

That is necessary for me to grasp the context of Scripture.

What Changed Today

Today’s change went into dd20cb2.

dd20cb2  update: blog writing templates

The core changes are these:

_project/blog-system/codex-feedback-prompt.md
  - Changed the prompt so every Bible verse includes both `Passage context` and `Reason for citation`
  - Defined the role of passage context as explaining the scene, argument, or exhortation flow
  - Added rules for using the `daily-review-scripture-context` class

_project/blog-system/daily-review-template.md
  - Added a `Passage context` slot for each of GPT-5.5's three recommended Bible verses

_project/blog-system/README.md
  - Added passage context as a required rule for Daily Review Bible verses

_sass/custom/customOverride.scss
  - Added `.daily-review-scripture-context` styling

This change is not large.

But it makes the faith section of Daily Review a little more honest. Even if AI recommends a verse, it now has to first confirm where that verse sits inside Scripture.

Result

Now GPT-5.5’s recommended Bible verses in Daily Review are divided into three parts.

Scripture text
Passage context
Reason for citation

I like this structure.

The Scripture text is the standard. Passage context is the place where that standard sits. The reason for citation is the connection to today’s record.

These three need to be separated.

Otherwise, GPT-5.5’s recommended Bible verses easily become a collection of good words. But if these three are separated, it at least makes me pause once before consuming the verse.

Today’s edit added one small field.

But the direction is clear.

Daily Review is not writing for the sake of comforting myself. It is a record for closing today honestly before God. Bible verses are the same. They should not be sentences that make me feel better, but words read within the passage that reflect today’s life.

That is why I added Passage context.