# mcp-keyward

> Use this tool when you need to securely manage API keys for HTTP requests, keeping them encrypted and invisible to large language models (LLMs). It solves the problem of protecting sensitive credentials while still allowing transparent injection into requests. The mcp-keyward tool takes encrypted API keys as input and outputs secure HTTP requests with the keys injected by the server.

Canonical page: https://skillsregistry.net/skills/ggc180820-mcp-keyguard  
JSON: https://api.skillsregistry.net/v1/skills/ggc180820-mcp-keyguard

## Description

Local encrypted vault for API keys. LLMs never see your real keys — the server injects them transparently into HTTP requests.

## Trust

- **Trust score (0–1):** 0.95
- **Verification tier:** verified
- **Last scanned:** 2026-09-28

## Facts

- **Version:** 1.0.0
- **Skill type:** atomic
- **Execution layer:** mcp-remote
- **Runtime environment:** api
- **Category:** ai-ml
- **Updated:** 2026-09-28

## Source

- **Source listing:** [Glama](https://glama.ai/mcp/servers/v156bqgzsm)
- **Repository:** <https://github.com/ggc180820/mcp-keyguard>

## Use it

Resolve this record through the SkillsRegistry MCP server (no auth, read-only):

```
claude mcp add --transport http --scope user skillsregistry https://api.skillsregistry.net/mcp
```

```json
{
  "jsonrpc": "2.0",
  "id": 1,
  "method": "tools/call",
  "params": {
    "name": "get_skill",
    "arguments": {
      "slug": "ggc180820-mcp-keyguard"
    }
  }
}
```

REST: `GET https://api.skillsregistry.net/v1/skills/ggc180820-mcp-keyguard` · pull for local use: `GET https://api.skillsregistry.net/v1/skills/ggc180820-mcp-keyguard/pull`

---
SkillsRegistry indexes agent skills from public registries and GitHub. Skills we have analysed are scanned with Circle-IR and scored on six dimensions; each listing states its scan coverage. More: https://skillsregistry.net/llms.txt