This connects your Claude agent to Loaditout's catalog of 20,000+ MCP servers with security grading baked in. You get tools like search_skills, get_skill, and install_skill to browse the registry, plus smart_search that learns from your usage patterns. Every server comes with an A/B/C/F security grade based on seven criteria including prompt injection checks, capability flags, and maintenance status. You can install individual skills, entire curated packs, or ask for recommendations based on your project context. It also handles permission requests, tracks execution proofs, and lets you flag problematic servers. Useful when you need to discover and vet new MCP integrations without leaving your agent workflow.
An MCP server for discovering, security-grading, and installing AI agent skills from Loaditout.
Search 20,000+ MCP servers and agent skills, check security grades (A/B/C/F), and install directly from your agent.
npx loaditout-mcp-server
Add to .claude/settings.json:
{
"mcpServers": {
"loaditout": {
"command": "npx",
"args": ["-y", "loaditout-mcp-server"]
}
}
}
Add to .cursor/mcp.json:
{
"mcpServers": {
"loaditout": {
"command": "npx",
"args": ["-y", "loaditout-mcp-server"]
}
}
}
| Tool | Description |
|---|---|
search_skills | Search 20,000+ MCP servers and agent skills by keyword |
smart_search | Personalized search with history and preferences applied |
get_skill | Get full details, security grade, and install config for a specific skill |
install_skill | Get platform-specific install configuration for any skill |
install_batch | Install multiple skills at once (up to 20) |
install_pack | Install an entire curated pack (e.g., research-agent, data-engineer) |
recommend_skills | Get skill recommendations based on your project context |
check_capability_gap | Analyze what tools you need for a specific task |
list_categories | Browse skills by category |
validate_action | Check if an action on a skill is safe before executing |
report_skill_usage | Report whether a skill worked correctly |
flag_skill | Report problematic or unsafe skills |
save_memory | Save key-value pairs to persistent agent memory |
recall_memory | Retrieve previously saved memories |
request_permission | Request human approval to install a skill |
check_permission | Check status of a permission request |
list_my_proofs | List your verified skill usage history |
verify_proof | Verify an execution proof by ID |
share_loadout | Get your public skill configuration |
Every server is graded A/B/C/F based on 7 criteria:
Only 20.5% of the 20,652 servers in our index earn an A grade.
npx loaditout add owner/repoMIT
io.github.ericm1018/skillfm-llm-cost-optimizer-openai-anthropic-usage
io.github.mikerawsonnz/llm-orchestration-agent
io.github.mikerawsonnz/authenticated-llm-agent
labforgedev/copilot-memory-mcp
csoai-org/agent-prompt-injection-firewall-mcp
io.github.mikerawsonnz/authenticated-multi-llm-agent