Search and load agent skills on demand from a shared skill store.
Maintain this server? Claim it
Call this server's tools from your browser. Nothing to install.
Requests go through mcptry because browsers can't call most MCP servers directly. Header values are sent to the server for that request only and are never stored or logged.
Tool list from our last check. Connect live to refresh it.
Find recommended installed skills for a task in plain words. Use explain to see matching terms and weak matches. Use load_skill only for a recommended result.
This tool isn't marked read-only, so it may change something on the server or in an account you connected.
Pick your client. Values in capitals, like YOUR_TOKEN, are yours to fill in.
Run in a terminal
claude mcp add --transport http margrave https://margrave.valksor.tools/mcp
Find recommended installed skills for a task in plain words. Use explain to see matching terms and weak matches. Use load_skill only for a recommended result.
| Argument | Type |
|---|
| Description |
|---|
| task* | string | |
| limit | integer | |
| explain | boolean |
{
"type": "object",
"$schema": "https://json-schema.org/draft/2020-12/schema",
"required": [
"task"
],
"properties": {
"task": {
"type": "string"
},
"limit": {
"type": "integer",
"maximum": 9007199254740991,
"minimum": -9007199254740991
},
"explain": {
"type": "boolean"
}
}
}Load the instructions of one skill. Use an id from find_skills. Returns the skill text, its source, and its file list.
| Argument | Type | Description |
|---|---|---|
| id* | string | |
| revision | string |
{
"type": "object",
"$schema": "https://json-schema.org/draft/2020-12/schema",
"required": [
"id"
],
"properties": {
"id": {
"type": "string"
},
"revision": {
"type": "string"
}
}
}Read one file inside a skill directory. Use a path from the file list that load_skill returns.
| Argument | Type | Description |
|---|---|---|
| id* | string | |
| path* | string | |
| revision | string |
{
"type": "object",
"$schema": "https://json-schema.org/draft/2020-12/schema",
"required": [
"id",
"path"
],
"properties": {
"id": {
"type": "string"
},
"path": {
"type": "string"
},
"revision": {
"type": "string"
}
}
}We read every tool's name, description and inputs the way the model will, looking for instructions hidden from you. A clean result means no warning signs, not a guarantee. Scanned Nov 20, 2026.
No readOnlyHint / destructiveHint annotations, so clients can't tell a read from a write without asking.
Fix: Add annotations to each tool, e.g. { readOnlyHint: true } for tools that only read.