Ontology XGJK

Typed knowledge graph for structured agent memory and composable skills. Use when creating/querying entities (Person, Project, Task, Event, Document), linkin...

This skill appears coherent for local structured memory: it reads/writes an append-only graph file under memory/ontology and provides Python utilities to manage entities and relations. Before installing, review the full scripts/ontology.py (the provided snippet was truncated) to ensure there are no hidden network calls or code that reads arbitrary system files. Consider these practical precautions: (1) run the tool in a restricted workspace or container so the graph file cannot reference sensitive host paths, (2) enforce file permissions on memory/ontology to prevent other processes from reading stored references, (3) avoid putting secrets directly into the graph (use secret_ref patterns and an appropriate secret store), and (4) if other skills are allowed to read this ontology, vet those skills because cross-skill access can leak sensitive references. If you want stronger assurance, provide the complete untruncated script for review or run the tool in an isolated environment first.

Type: OpenClaw Skill Name: ontology-xgjk Version: 1.0.4 The skill bundle provides a structured knowledge graph (ontology) system for agent memory and state management. The core logic in `scripts/ontology.py` includes security-conscious features such as path traversal protection via `resolve_safe_path` and schema constraints that explicitly forbid the direct storage of secrets (passwords/tokens) in the `Credential` type. The behavior is entirely consistent with the stated purpose of managing entities and relations within the local workspace.