Production-ready prompt management for LLM agent swarms. Rust 2024 Edition.

• Vibe Coding: Natural language to deliverable for non-technical users
• Semantic Search: FAST (FTS5) + SMART (ONNX embeddings) + Hybrid
• Version Control: Full semver lineage with diff and rollback
• Security: Prompt injection detection, RBAC, audit logging
• Swarm Integration: Role-based bundles, handoff templates
• Enterprise: OpenTelemetry, health checks, graceful shutdown

# Install CLI
cargo install prompthub

# Initialize database
prompthub init

# Add a prompt interactively
prompthub add --interactive

# Search prompts
prompthub search "error handling" --mode hybrid

# Generate shell completions
prompthub completions bash > /etc/bash_completion.d/prompthub

use prompt_hub::{PromptHub, HubConfig};
use std::path::Path;

#[tokio::main]
async fn main() -> Result<(), Box<dyn std::error::Error>> {
    let config = HubConfig::default();
    let hub = PromptHub::new(Path::new("prompthub.db"), config).await?;

    // Register, search, vibe_code, etc.
    Ok(())
}

# Run all tests
cargo test --all-features

# Run benchmarks
cargo bench

# Build documentation
cargo doc --all-features --no-deps

# Run with specific features
cargo run --bin prompthub --features tui -- tui

docker build -f docker/Dockerfile -t prompthub .
docker-compose -f docker/docker-compose.yml up

The project includes an automation that triggers updates to TODO.md whenever a new audit report is added to docs/audits/.

• GitHub Action: .github/workflows/audit_sync.yml automates this in CI.
• Local Watcher: scripts/audit_watcher.sh can be used for local development.

See docs/architecture.md for C4 model diagrams and system design.

MIT OR Apache-2.0