graphonomous/docs/skills

Documents

• Attention Engine — Autonomous focus management — rank goals by urgency, knowledge gaps, and topology complexity.
• Belief Revision — AGM-style belief management — expand, revise, or contract beliefs with automatic contradiction detection and confidence propagation.
• Consolidation — Trigger graph maintenance — decay weak knowledge, prune, merge duplicates, strengthen relationships.
• Coverage and Review — Evaluate whether you know enough to act on a goal, or need to learn more.
• Epistemic Frontier — Uncertainty-Guided Investigation — Find nodes where investigation would most reduce graph uncertainty.
• Evidence Path Tracing — Find the lowest-cost evidence path between two knowledge nodes using weighted Dijkstra with optional K-shortest alternate paths via Yen's algorithm.
• Filesystem Sync — Batch Ingest to Graphonomous — Sync filesystem changes into the Graphonomous knowledge graph.
• Goal Management — Track durable intent across sessions — create goals, update progress, link knowledge.
• Graph Health Check — Combined diagnostic that assesses knowledge graph quality in one pass.
• Graph Inspection — Read-only inspection of the knowledge graph — no side effects.
• Graphonomous Bootstrap — Initialize the Graphonomous-first memory loop for this session.
• Graphonomous MCP — Agent Skills — Graphonomous is a continual learning engine exposed as 5 MCP machines. It maintains a self-evolving knowledge graph of semantic, procedural, episodic, temporal, outcome, and goa…
• Graphonomous Workflows — End-to-end recipes for common tasks. Pick the workflow that matches your situation.
• Intentional Forgetting — Structured Knowledge Removal — Four forgetting modes plus budget-aware policy pruning.
• Learning Loop — Close the feedback loop so the graph learns what works and what doesn't.
• Retrieve and Remember — Search the knowledge graph and store new knowledge.
• Store Knowledge — Save a specific piece of knowledge to the Graphonomous knowledge graph.
• Topology and Deliberation — Detect and reason through circular dependencies (cycles) in knowledge.
• Watch Directory — Continuous Filesystem Sync — Start a polling loop that detects filesystem changes and ingests them into Graphonomous.