Agent skill

LLM coding agents can drive ntropy well, since the CLI is fully scriptable, but they have to know the house rules. The repository ships an agent skill that teaches them.

Install

The quickest install is the skills CLI, which places the skill into Claude Code, Cursor, Codex, and most other SKILL.md-aware agents. It installs into the project by default and for your user with -g:

npx skills add jakobwesthoff/ntropy

Or skip the tooling and copy the directory by hand. For Claude Code it belongs at ~/.claude/skills/ntropy (every project) or <project>/.claude/skills/ntropy (just that project):

git clone https://github.com/jakobwesthoff/ntropy.git
cp -R ntropy/skills/ntropy ~/.claude/skills/ntropy

The skill's description marks it as relevant to any task involving ntropy, a note vault, or .ntropy-vault files, so an agent with the skill installed picks it up on its own the moment a task touches one.

What the skill teaches

skills/ntropy/ holds a SKILL.md with the vault model, rules for agents, a do/don't table, and a command reference. Beside it are reference docs on writing notes, querying, vaults, views, the website export with its navigation settings, site themes, and document themes.

The rules are the part an agent gets wrong on its own:

  • Always pass -n, and --print on new and today. Without them ntropy opens a picker or the editor and blocks whenever a terminal exists; capturing the output does not prevent that.
  • Never hand-create files in all-notes/. ntropy allocates the ULID, the location, and the file name; only the content is the agent's.
  • Author a note with ntropy new --empty -p <title> followed by ntropy write <path>, feeding the whole note on stdin.
  • Run ntropy reconcile after editing a note file directly instead of through write.
  • Check the active vault with ntropy info before changing anything, never write into the view directories, and never store id or dates in frontmatter.