deep-reflect

What problem does it solve?

Deep-reflect turns ad-hoc insights and findings into evidence-backed reflections by automatically retrieving relevant artifacts, running iterative multi-sample analyses, and surfacing structurally meaningful signals (tensions, gaps, convergences, dependencies) so teams can make decisions with grounded context instead of intuition.

Core Features & Use Cases

• Evidence-driven cross-referencing: Retrieves candidate artifacts from the organization's memory and inspects file content to justify each signal with excerpts and provenance.
• Multi-mode analysis: Supports deep (multi-sample consensus), focused (topic-filtered), and quick (single-pass) modes to trade off latency and depth.
• Signal ontology & graph integration: Classifies relationships (tension, convergence, gap, dependency, etc.), writes a reflective artifact to memory, and creates graph edges to record reasoning and consequences.
• Use Case: Use after a design decision or incident to understand whether the insight contradicts prior decisions, uncovers missing prerequisites, or reinforces an existing pattern.

Quick Start

Ask the assistant to "deep-reflect 'finding: Neo4j queries are slow'" to run an evidence-based cross-reference against the team's memory.