Rusty Page Indexer Skill

This skill allows agents to index and query local documentation using rusty-page-indexer. Unlike standard vector databases, it focuses on Structural Context, preserving the document's hierarchical tree for more logical retrieval.

Why Structure Matters

Standard RAG often slices documents into arbitrary chunks, losing context. rusty-page-indexer maintains a Folder → File → Section hierarchy.

• Logical Retrieval: Instead of just matching a sentence, you match a logical section (e.g., an entire API endpoint description).
• Project Awareness: The LLM understands where a file lives (e.g., src/auth) giving it contextual cues about the code's purpose.

Key Capabilities

• Multi-repo indexing: Index multiple projects independently
• Cross-repo search: Query across all indexed repositories
• Parallel processing: Fast indexing with Rayon
• Unified tree structure: Folder -> File -> Section hierarchy
• Model flexibility: Works with OpenAI, Ollama, and OpenAI-compatible APIs

Commands

1. Install Tool

Install the rusty-page-indexer binary (Linux/macOS).

curl -fsSL https://raw.githubusercontent.com/Algiras/rusty-pageindex/main/install.sh | bash

2. Configure Authentication

OpenAI (remote):

rusty-page-indexer auth --api-key "sk-..." --model "gpt-4o-mini"

Ollama (local):

rusty-page-indexer auth --api-key "ollama" --api-base "http://localhost:11434/v1" --model "llama3.2"

Other OpenAI-compatible APIs:

# Groq
rusty-page-indexer auth --api-key "your-key" --api-base "https://api.groq.com/openai/v1" --model "llama3-70b-8192"

# Together AI
rusty-page-indexer auth --api-key "your-key" --api-base "https://api.together.xyz/v1" --model "meta-llama/Llama-3-70b-chat-hf"

3. Index Documents

Recursively index a directory of documents (.md, .pdf).

# Fast index (no LLM calls, instant)
rusty-page-indexer index <path>

# With LLM-generated summaries (higher quality, slower)
rusty-page-indexer index <path> --enrich

# Two-Phase Strategy (Recommended)
# 1. Index the whole folder quickly for structure
rusty-page-indexer index ./projects
# 2. Enrich only the critical "Holy Grail" docs (specs/manuals)
rusty-page-indexer index ./projects/docs/architecture_spec.md --enrich

[!TIP] Use --enrich sparingly on large folders. Index generally first, then enrich specific key files for deep reasoning.

### 4. List Indexed Repositories
View all indexed repositories with their sizes and paths.

```bash
rusty-page-indexer list

5. Query Index

Perform semantic search over indexed documents.

# Search all repositories
rusty-page-indexer query "How does authentication work"

# Path-Scoped Query (Search only specific component/folder)
# Useful for multi-repo work or filtering to 'backend' only.
rusty-page-indexer query "auth" --path backend

# Override model for this query (using argument separator)
rusty-page-indexer query "Explain API" -- --model gpt-4o

6. Clean Up Indices

Remove indexed repositories.

# Remove specific index
rusty-page-indexer clean <repo-name>

# Remove all indices
rusty-page-indexer clean --all

7. Check Status

Verify installation and configuration.

rusty-page-indexer info

Model Compatibility

OpenAI Models (Remote)

Local Models (Ollama)

Troubleshooting

The tool provides detailed error messages:

Typical Workflow

# 1. Install (if needed)
curl -fsSL https://raw.githubusercontent.com/Algiras/rusty-pageindex/main/install.sh | bash

# 2. Configure for Ollama (free, local)
rusty-page-indexer auth --api-key "ollama" --api-base "http://localhost:11434/v1" --model "llama3.2"

# 3. Index your project
rusty-page-indexer index ./docs --enrich

# 4. Search
rusty-page-indexer query "how does authentication work"

# 5. View indexed repos
rusty-pag

...