Build Index

Generate INDEX.md (Compact Retrieval Index) for an existing documentation package.

Usage

/build-index prisma@latest         # Generate index
/build-index prisma@latest --force  # Regenerate even if exists

Input

$ARGUMENTS format: {config_name}@{version} (e.g., prisma@latest, nextjs@16.1.3)

Optional flags:

• --force - Regenerate even if INDEX.md already exists

Prerequisites

The contents/ directory and olore-lock.json must already exist at vault/packages/{config_name}/{version}/. Run /download-docs first if they don't.

test -d vault/packages/{config_name}/{version}/contents && echo "OK" || echo "NOT_FOUND"

If contents/ not found:

error: {config_name}@{version} contents/ not found - run /download-docs first

Execution Steps

Step 1: Parse Arguments and Load Metadata

Parse $ARGUMENTS:

• Extract config_name and version from {config_name}@{version}
• Check for --force flag

Read metadata:

cat vault/packages/{config_name}/{version}/olore-lock.json

Step 2: Check if Already Built

test -f vault/packages/{config_name}/{version}/INDEX.md && echo "EXISTS" || echo "NOT_FOUND"

If exists and no --force flag:

skip: {config_name}@{version} INDEX.md already built (use --force to rebuild)

Return early.

Step 3: Determine Tier

file_count=$(find vault/packages/{config_name}/{version}/contents -type f \( -name "*.md" -o -name "*.mdx" \) | wc -l)
total_size=$(du -sk vault/packages/{config_name}/{version}/contents | cut -f1)

Tier	Criteria
1	< 30 files AND < 500KB
2	30-100 files OR 500KB-2MB
3	> 100 files OR > 2MB

Step 4: Generate INDEX.md

Read the appropriate index template based on tier:

# Tier 1
cat vault/packages/docs-packager/1.0.0/templates/index-tier1.md

# Tier 2
cat vault/packages/docs-packager/1.0.0/templates/index-tier2.md

# Tier 3
cat vault/packages/docs-packager/1.0.0/templates/index-tier3.md

Create vault/packages/{config_name}/{version}/INDEX.md following the template structure.

Format: Compact @section lines consumed by olore inject:

@section:keyword1,keyword2=path/to/file.mdx;keyword3,keyword4=path/to/dir/

Delimiter	Separates
@	Section name from entries
;	Entries within a section
=	Keywords from file path
,	Keywords

• Paths are relative to contents/ (do NOT include the contents/ prefix)
• Paths ending in / are directories; paths ending in .md/.mdx are files
• Keywords should be actual API names, method names, config keys — not descriptions

Process:

1. Read each file in contents/
2. Extract keywords: API names, method names, config keys, class names, function signatures
3. Group entries into logical @section lines (one line per section)
4. Write INDEX.md following the tier template

Step 5: Return Summary

Return ONLY a brief summary:

done: {config_name}@{version} INDEX.md built, {entry_count} entries, tier {tier}

Outputs

• vault/packages/{config_name}/{version}/INDEX.md - Compact Retrieval Index