SkillFoxx

Ollama Buddy skill

Emacs package for Ollama and cloud models with a skill library that can be exported for Claude Code.

  • 104 ★ on the repository
  • Last commit
  • GPL-3.0
Skill

Extended access

Why this level

  • The model gets tools to read and write files and run commands, and the unrestricted mode removes confirmations.
All reasons and checks

captainflasmr/ollama-buddyAuthor:captainflasmr

Install

In your terminal, with SkillFoxx CLI

npx skillfoxx add skills/ollama-buddy

Detects the agents on your machine, shows the access level and pins the version.

Other ways to install

Access and checks

How we review

Extended accessThe tool runs code or makes network calls and only reads external data.

  • The model gets tools to read and write files and run commands, and the unrestricted mode removes confirmations.

This is third-party code. Review the repository files before installing.

What it does

Ollama Buddy is an Emacs package for version 29.1 and newer for working with local Ollama models and remote providers: OpenAI, Claude, Gemini, Grok, DeepSeek, OpenRouter and any OpenAI-compatible server. Chat, in-buffer rewriting, attachments, sessions and history are reachable from `transient` menus. The model can call Emacs functions: read and write files, search buffers, run shell commands. Safe mode keeps read-only tools. The prompt library includes skills with developer, writer and tutor roles, invoked through `/skill` or `@skill(name)`. `M-x ollama-buddy-user-prompts-export-skills` exports them to `~/.claude/skills/` as `SKILL.md` files.

Who it is for. Emacs users who keep models local and want one setup for both chat and Claude Code skills.

Good fit when

  • A developer rewrites a paragraph or function in an Emacs buffer without opening a chat window.
  • Answers from a local Ollama model and a cloud provider must be compared on one prompt.
  • A set of skills for code review, documentation and release notes has to move into `~/.claude/skills/`.
  • Work happens in `org-mode` and a task plan is needed as TODO headings, which the `/plan` command produces.

Not a fit when

  • The editor is not Emacs; the package does not run without it.
  • Project policy forbids the model from running shell commands.
  • A single ready skill for one task is needed; here they ship as a shared library rather than separate packages.

Example request

Export all Ollama Buddy skills to `~/.claude/skills/` and show which of them relate to code review.

Limitations

Emacs 29.1 or newer is required. Cloud providers use your own keys, and local mode needs a running `ollama serve`. The unrestricted mode (`☠`) removes every confirmation prompt for tool calls, shell commands included. The repository skills are written for Emacs, `org-mode` and the package itself, so only some suit general Claude Code work. License: GPL-3.0. Detailed guides to roles and menus live in a separate full README, and video walkthroughs sit on the author channel.

How to disable. Remove `use-package` from the Emacs configuration and delete the package. Delete exported skills from `~/.claude/skills/` and saved sessions separately.

FAQ

How do I invoke a skill inside chat?

Use the `/skill` command or write `@skill(name)` in the prompt; the entry expands to the skill content.

How do I export skills for Claude Code?

Run `M-x ollama-buddy-user-prompts-export-skills`. Each skill becomes a directory with `SKILL.md` under `~/.claude/skills/`, and with a `C-u` prefix the command asks for other directories.

What does safe mode do?

It limits tools to read-only. The unrestricted mode removes confirmation prompts, so enable it deliberately.

Which providers are supported?

Local and cloud Ollama, OpenAI, Claude, Gemini, Grok, GitHub Copilot, Codestral, DeepSeek, OpenRouter, OpenCode Go and OpenAI-compatible servers.

How do I start a local model?

Run `ollama serve`, open Emacs and press `C-c o` if you bound the README keys. Pick `[o]` to open chat.

Discussion

No questions or reviews yet.

Skills

oterm is a terminal chat for Ollama, OpenAI, Anthropic and other providers: SQLite-stored sessions, MCP servers, built-in and custom tools

CLIBroad access2.4K
Official

Ollama runs open models on your own machine and serves them to agents through a local API compatible with OpenAI and Anthropic.

CLIExtended access186.9K

A Claude Code workspace folder: hooks with a local Ollama detect the project type, pick skills and remind you to update the docs

WorkflowExtended access307

A CLI and Python library for OpenAI, Claude, Gemini, local and other models: prompts, chats, tools, embeddings and a SQLite log

CLIExtended access12.8K

An MCP server that hands bounded tasks to a local model or a cheap cloud API, while Claude stays the architect and saves tokens

MCP serverExtended access122

README in short

Basic setup fits one `use-package` block. Detailed tutorials, roles and menus live in a separate full README and video walkthroughs.

README badge

Report a problem