Personal Ontology

What to know before installing: - Privacy: The skill is designed to scan your local notes/journal (including archived notes) to extract beliefs, predictions, goals, and projects. If you allow bootstrap or enable daily/automated scans, the agent will read those files. Only enable scanning you are comfortable with and review candidates before committing. - Local writes: When you accept suggestions the skill will create/update markdown files in My_Personal_Ontology and may write small local state files (e.g., memory/ontology-nudges.json). Back up important notes first if you are cautious. - No network credentials requested: The skill does not ask for API keys or other credentials. If you later integrate it with external task boards or services, treat those integrations as separate and only supply credentials if you trust the code/service. - Optional dependency: The renderer script can use 'beautiful-mermaid' for ASCII/SVG export; that package is not installed automatically. If you use that feature, install it from the public registry or inspect it first. - Safety steps: (1) Run the bootstrap on a small subset of notes or a safe test vault to see what gets extracted. (2) Keep automated/passive scans disabled until you trust the suggestions and behavior. (3) Inspect scripts/render-ontology.js (present) before executing and set ONTOLOGY_DIR to a dedicated folder. (4) If you share the agent or machine, be aware the ontology files may contain sensitive personal content. Overall: the skill is internally consistent with its stated purpose. The main risks are privacy/accidental exposure from scanning and the fact it will write files locally — these are expected for a notes-scanning ontology tool. If you want stronger guarantees, run it in a sandboxed workspace or review candidate outputs before committing.