SkillAgentSearch skills...

savile

A high-fidelity, local-first MCP server and protocol for storing, versioning, syncing, and evaluating AI agent prompts and skills

Install / Use

claude mcp add mr8lu -- npx -y github:mr8lu/savile

If the server publishes to npm under a different name, use that package instead — check the repo README.

About this skill
🔌

MCP Server

Model Context Protocol server

Quality Score

78/100

Supported Platforms

Claude Code
Claude Desktop

Tags

Our assessment of savile

savile scores 78/100 on our quality scale, 3702nd of 4,700 Development & Engineering skills we index.

Its MCP Server is 4.5 KB long, well organised into 12 sections with 9 code examples: a solid amount of guidance for an agent.

It has 3 GitHub stars, so there is little community track record yet; judge it on its content.

Substance
26/30
Structure
20/20
Description
15/15
Adoption
3/20
Freshness
15/15

Maintenance, license and trust

  • The repository was last updated today, so savile is actively maintained.
  • Our last check on 2026-09-23 found the source still online.
  • It is released under the MIT license, a permissive license that allows use, modification and commercial use with attribution.
  • Its trust signals score 92/100, with 1 caution from licensing, adoption, age or documentation. These come from repository metadata, not a code audit — read the skill file before letting an agent act on it.

Safety scan

No issues found

Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful.

AI review by kimi-k2.7-code on 2026-10-03. Automated pattern scan on 2026-10-03. It catches known dangerous patterns, not every risk — read a skill before letting an agent act on it.

savile compared with similar skills

All 4 of these similar skills score higher than savile; compare them before choosing.

SkillScoreStarsUpdatedFormat
savile (this skill)by mr8lu783todayMCP Server
taste-skillby Leonxlnx100116.8k8d agoCLAUDE.md
Agent-Reachby Panniantong100112.3k8d agoCLAUDE.md
rufloby ruvnet10081.9ktodayCLAUDE.md
headroomby headroomlabs-ai10079.8ktodayCLAUDE.md

Frequently asked questions

How do I install savile?
Run claude mcp add mr8lu -- npx -y github:mr8lu/savile. The install tabs above show the steps for each supported agent.
Which AI agents does savile work with?
It is written for Claude Code and Claude Desktop, as a MCP Server file. Other agents that read the same format can often use it too.
Is savile safe to use?
Our scan of the whole file found no instruction hijacking, hidden characters, credential access, data exfiltration or destructive commands. An AI review of the same text found nothing harmful. It is MIT-licensed and scores 92/100 on trust signals. Skills are instructions an agent will follow, so read the file before installing it and do not approve commands you do not understand.
Is savile still maintained?
The repository was last updated today, so savile is actively maintained.

SAVILE: Single Source of Truth Engine for Agent Prompts & Skills

Python 3.11+ Model Context Protocol License: MIT

SAVILE acts as the definitive "Next.js for Agent Prompts." It provides a powerful local-first environment to write, version, and compile AI logic into universally portable formats.

It operates as a three-tier architecture that orchestrates your agent logic:

  1. The Methodology (Method Framework): The core psychology and workflow definitions behind how your agents think.
  2. The SSoT Vault (Savile): A Git-native database where you fork, customize, and version your specific personas, frameworks, and evaluations.
  3. The Universal Export (Vercel Skills & MCP): Savile seamlessly compiles your custom logic into portable Vercel Labs SKILL.md formats for global distribution via npx skills add, or serves them dynamically on localhost via the Model Context Protocol (MCP).

🧐 Why SAVILE?

Savile exists to end "prompt drift" and vendor lock-in. Instead of configuring instructions directly inside isolated UIs (like Cursor, Claude Desktop, or V0), you maintain them centrally as pure markdown files.

  • Fork, Modify, Serve: Treat your agent configurations like open-source software. Fork the baseline logic vault, adapt the QA and Dev personas to your company's standards, and deploy.
  • Universal Portability: Because Savile exports to the Vercel Labs Skills standard, a customized agent persona can be downloaded into any terminal environment instantly.
  • De-anthropomorphized Scale: Savile enforces strict, functional naming rules (ux-designer, quality-assurance), ending the chaotic practice of giving AI agents human names and standardizing your robotic workforce.

🚀 Quick Start

1. Installation

git clone https://github.com/mr8lu/savile.git
cd savile
uv sync

2. The Universal Lifecycle

Savile provides a streamlined lifecycle for managing your agent intelligence.

bootstrap: Initialize a new SSoT vault or clone an existing one from a remote source. Interactively links your core Method framework.

uv run savile bootstrap
# Or clone a team vault
uv run savile bootstrap --source git+ssh://github.com/user/my-logic-vault.git

import: Bring logic modules or skills from outside repositories into your local vault.

uv run savile import git+ssh://github.com/user/remote-logic.git

import-system: Import a custom skill/agent from the system-wide directory (~/.gemini or ~/.agents) or a custom directory into your local vault.

uv run savile import-system sprint-status
# Or specify a custom source directory
uv run savile import-system custom-skill --dir /path/to/custom_agents

create: Scaffold a new functional agent persona. (Savile strictly enforces non-human, functional naming conventions).

uv run savile create ux-designer --desc "Analyzes wireframes and user flows"

update: Refresh existing metadata for a persona or framework in your vault.

uv run savile update ux-designer

export: Compile your raw personas/ and frameworks/ down into the universally portable Vercel Labs Skill format (SKILL.md), injecting YAML frontmatter automatically.

# Outputs to dist/skills/
uv run savile export

serve: Boot the local MCP server to dynamically stream your logic vault into IDEs like Cursor, Windsurf, or Claude Desktop.

uv run savile serve

🛠️ Ecosystem Integrations

Vercel Labs Skills

By running savile export, you generate standard SKILL.md artifacts. If you commit these to a public GitHub repository (e.g., your fork of Savile) or deploy them via GitHub Pages, developers anywhere can import your customized agent by running:

npx skills add <your-username>/savile --skill ux-designer

Model Context Protocol (MCP)

Running savile serve exposes your local vault directly to your editor. Changes you make to personas/ux-designer.md are instantly available to Claude without requiring a restart.


📄 License

MIT License. See LICENSE for details.

Built with precision for the sovereign developer.

Related Skills

View on GitHub
GitHub Stars3
CategoryDevelopment
Updated11h ago
Forks0

Languages

Python

Trust signals

92/100

From repository metadata: license, adoption, age and documentation. Not a code audit — see the Safety scan above for what the skill file itself contains.

1 low