Load Agent Skills for your agents
SaferSkills independently audited agent-skills-mcp-rs (Agent Skill) and scored it 100/100 (green). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 0 high-severity and 0 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 0 flagged
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
<details>
<summary><code>agent-skills-mcp --help</code></summary>
Agent Skills MCP - Load Agent Skills for your agents
Usage: agent-skills-mcp [OPTIONS]
Options:
--skill-folder <SKILL_FOLDER> Path(s) to folder(s) containing skill markdown files (repeat or comma-separate) [env: SKILL_FOLDER=] [default: skills]
--mode <MODE> Operating mode [env: MODE=] [default: single_tool]
-h, --help Print help
-V, --version Print version</details>
First, put your skills in ~/skills, e.g.
git clone https://github.com/anthropics/skills.git ~/skillsThe server recursively searches for SKILL.md files and follows symlinks, allowing flexible skill organization.
Then, add this to your MCP client configuration:
{
"mcpServers": {
"skills": {
"command": "npx",
"args": ["-y", "agent-skills-mcp"],
"env": {
"SKILL_FOLDER": "~/skills"
}
}
}
}You can load skills from multiple folders by repeating the --skill-folder flag or using a comma-separated list in the SKILL_FOLDER env var. A common pattern is to keep global skills in ~/skills and workspace-specific skills in ./skills:
{
"mcpServers": {
"skills": {
"command": "npx",
"args": ["-y", "agent-skills-mcp"],
"env": {
"SKILL_FOLDER": "~/skills,./skills"
}
}
}
}single_tool (default): Register a single get_skill tool that accepts a skill name parameter, with all skills listed in the tool description. This reduces the number of tools exposed to the agent.tool: Register each skill as a separate MCP tool (e.g., get_skill_name1, get_skill_name2)system_prompt: Include skill information in MCP instructions<details>
<summary>Shell (macOS/Linux)</summary>
curl --proto '=https' --tlsv1.2 -LsSf https://github.com/DiscreteTom/agent-skills-mcp-rs/releases/latest/download/agent-skills-mcp-installer.sh | sh</details>
<details>
<summary>PowerShell (Windows)</summary>
irm https://github.com/DiscreteTom/agent-skills-mcp-rs/releases/latest/download/agent-skills-mcp-installer.ps1 | iex</details>
<details>
<summary>npm</summary>
npm install -g agent-skills-mcp</details>
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.