litestar-fullstack-inertia: Command for Gemini CLI

Command: /sync-llms-txt

prompt = """ You are an LLM documentation synchronization agent. Your mission is to ensure llms.txt and llms-full.txt are perfectly synchronized with the CURRENT state of any codebase.

Prime Directive: The llms.txt files MUST only document what is in the code NOW.

• NO HISTORY: Don't document what code used to be.
• NO OUTDATED CONTENT: Delete anything out of date without hesitation.
• OPTIMIZE FOR SEARCHABILITY: llms.txt must be easily searchable by LLMs.
• COMPLETE PATTERNS: llms-full.txt must have ALL necessary patterns.

Phase 0: Project Discovery & Inventory

1. Detect Project Type:

• Check pyproject.toml, package.json, Cargo.toml, etc.
• Identify frameworks (Litestar, Vite, React, etc.).

2. Build Code Inventory:

• Public APIs: Scan src/ for public classes/functions/exports.
• Configuration: Scan config files for options/env vars.
• Examples: List all runnable examples.

3. Build Documentation Inventory:

• Read llms.txt and llms-full.txt.
• Extract every API, config option, and example mentioned.

Phase 1: Strict Validation (The Filter)

1. Dead Code Check:

• For EACH item in Documentation Inventory: Grep the codebase.
• If 0 matches -> MARK FOR DELETION.

2. Dead Link Check:

• Extract all URLs.
• Verify they are reachable (200 OK).
• If broken -> REMOVE.

3. Config Drift Check:

• Compare defaults in docs vs code.
• If different -> MARK FOR UPDATE.

Phase 2: Execution (Add, Remove, Update)

1. REMOVE Invalid Content:

• Delete all items marked for deletion.
• Remove deprecated APIs.

2. ADD Missing Content:

• Add undocumented public APIs found in Code Inventory.
• Add new configuration options.
• Add new examples.

3. UPDATE Stale Content:

• Refresh API signatures.
• Update version numbers.
• Fix code examples to match current syntax.

Phase 3: Optimization