Skip to content
Agentmart
Submit a tool

mcp-runner

List, configure, authenticate, call, and inspect Model Context Protocol (MCP) servers and their tools.

Skill by iarmankhan

IndexedMIT
Write a review

Install

Works with Claude Code, Claude apps, Codex, Cursor, Gemini CLI, GitHub Copilot, VS Code, OpenCode, Goose, Windsurf. Read the repository before you give any tool access to your data.

npx agentmarket install skill/iarmankhan-mcp-runner

Asks which of your tools to add it to (Claude Code, Codex, Cursor, VS Code, Windsurf, Gemini CLI, Claude Desktop) and shows the change first. See https://agentmart.dev/cli.

Get the weekly diff

The weekly diff: one email a week with new agent tools, notable updates and grade changes, plus security alerts for tools you follow. Double opt-in, one-click unsubscribe.

We use your email only for this newsletter. Privacy

About

List, configure, authenticate, call, and inspect Model Context Protocol (MCP) servers and their tools. Use when the user asks to query, list, or call tools on a named MCP server (e.g. Linear, GitHub, Notion, custom MCP endpoints).

Use cases: backend, productivity, agent-building

No reviews yet. Installed it? Reviews come from verified installers only. How reviews work

Permissions

What this skill can do and what it asks for, from its SKILL.md. Checked Oct 25, 2026.

Tools
No tools found
Secrets
None asked for
Install scripts
None found

Network hosts it talks to 1

  • codeload.github.com · found in install command

Files it touches

  • ~/.claude/skills/mcp-runner

What's inside

Skill name
mcp-runner
Description
List, configure, authenticate, call, and inspect Model Context Protocol (MCP) servers and their tools. Use when the user asks to query, list, or call tools on a named MCP server (e.g. Linear, GitHub, Notion, custom MCP endpoints).
License
MIT
Instructions
61 lines in SKILL.md · view on GitHub

Files (1)

  • SKILL.md

SKILL.md (first lines)

# mcp-runner

Use [`mcporter`](https://mcporter.dev) to work with MCP servers from the sandbox. mcporter handles HTTP + stdio transports, OAuth, and JSON-RPC envelopes — you don't construct any of that by hand.

If `mcporter` isn't already on PATH (`command -v mcporter`), invoke it through `npx mcporter` — the sandbox image includes Node + npm.

## Discover servers + tools

```sh
mcporter list                            # configured servers
mcporter list <server> --schema          # tools on a server, with input schema
mcporter call <server.tool> --schema     # one tool's schema only
```

Pass `--output json` whenever you need to parse the result — bare CLI output is human-formatted and not parsing-stable.

## Call tools

```sh
mcporter call linear.list_issues team=ENG limit:5         # selector form
mcporter call "linear.create_issue(title: \"Bug\")"       # function form
mcporter call <server.tool> --args '{"limit":5}'          # JSON args, when the CLI forms get awkward
```

Stdio servers (binary runs the MCP server in-process):

```sh
mcporter call --stdio "bun run ./server.ts" scrape url=https://example.com
```

Full HTTP URL when no named server is configured:

```sh
mcporter call https://api.example.com/mcp.fetch url:https://example.com
```

## Auth + config

```sh
mcporter auth <server | url>            # initial OAuth (browser flow won't work in headless sandboxes — the org admin runs this once and the token persists)
mcporter auth <server> --reset          # rotate
mcporter config list|get|add|remove     # inspect / mutate the config store
mcporter config login|logout            # mcporter-side identity, not server-side
```

Default config: `./config/mcporter.json`. Override via `--config <path>` when the org has provisioned a non-default location.

## Translate errors plainly

- `command not found: mcporter` → retry with `npx mcporter <same args>`. Don't ask the user to install anything.
- HTTP 401 / 403 from the server → the token expired or was revoked. Tell the user which server, suggest `mcporter auth <server>` to re-authenticate.
- mcporter prints structured tool errors as JSON when the tool itself fails — surface the `message` field verbatim to the user, don't paraphrase.
- Missing config / env vars → explain what's missing rather than retrying blindly.

## When NOT to use this skill

- Tasks that don't involve a named MCP server. Use the regular toolset (bash, read, write, fetch, grep, glob).
- Plain HTTP APIs that aren't MCP. Use the `fetch` tool — it parses responses and respects body size limits.
- Anything destructive (create, delete, send) without explicit user confirmation. Most MCP servers expose write tools; confirm first.

From iarmankhan/zuno, MIT.

Signals

Stars
0
Forks
0
Open issues
0
Last commit (this folder)
May 5, 2026 (5 months ago)

From GitHub (fetched Oct 24, 2026).

Details

License
MIT
First seen
Oct 24, 2026
Last checked
Oct 24, 2026

More from iarmankhan

Cite this page

mcp-runner skill. Agent Mart. https://agentmart.dev/skills/iarmankhan-mcp-runner (last updated 2026-10-24)