deep-recon

What problem does it solve?

Orchestrates a multi-agent reconnaissance workflow inside a user's knowledge base, parsing input, dispatching subagents, cross-pollinating findings across rounds, and producing a structured recon document.

Core Features & Use Cases

• Coordinated agent orchestration: four specialized agents (Explorer, Associator, Critic, Synthesizer) run in parallel rounds to map a topic from multiple angles.
• Interactive and autonomous modes: supports live steering between rounds or fully automated execution with a final recon.
• Obsidian-native output & traceability: generates a final markdown recon with a process log, plus per-round agent reports and sources.
• Vault and web synthesis: reads vault content and can fetch web sources to build a comprehensive perspective.
• Extensible workflow: optional pdf collection and structured prompts for reproducible recon sessions.

Quick Start

Run a deep recon on a topic; the system will dispatch four subagents, gather vault and web material, and output an Obsidian-ready recon with a process log.