Getting it into your agent
One page per mod, every tool's command on it. A separate URL per tool would split the same page into five that compete with each other.
git clone --depth 1 https://github.com/alexmmatos/arthur-mcpWrote this? Show the measurements
A badge with what this costs and how it scanned, read live from this page, so it follows the numbers instead of freezing them. Markdown for a README, HTML for a documentation site or a project page.
[](https://agentmods.dev/agents/alexmmatos/arthur-mcp/tool-instructor)<a href="https://agentmods.dev/agents/alexmmatos/arthur-mcp/tool-instructor"><img src="https://agentmods.dev/badge/agents/alexmmatos/arthur-mcp/tool-instructor.svg" alt="Measured on agentmods" height="20"></a>What it costs to keep this loaded
Counted locally with the o200k_base tokenizer, which is exact for GPT models; Claude uses its own tokenizer and its counts differ. Treat this as one consistent yardstick across the catalogue rather than a bill. Prices are per million input tokens.
| Model | Per session | Once invoked |
|---|---|---|
| Fable 5.1 | $0.00069 | $0.01546 |
| Opus 5 | $0.00034 | $0.00773 |
| Sonnet 5 | $0.00014 | $0.00309 |
| Haiku 4.5 | $0.00007 | $0.00155 |
Grade A, and why
tool-instructor scanned grade A with 0 findings against 26 rules in 11 categories — prompt injection, anti-refusal, data exfiltration, privilege escalation, supply chain, agent snooping, system-prompt leakage, SSRF and excessive agency — measured 4d ago.
A static scan of the body, not an audit. Every finding is printed with the line that produced it so you can judge whether it matters here. A mod is markdown that instructs an agent; that is exactly why what it instructs is worth reading.
Nothing flagged
None of the 26 patterns this scan looks for appear in this file: no shell pipes, no recursive deletes, no credential paths, no hidden text, no instruction-override or anti-refusal phrasing, no agent-config snooping. That is not a guarantee, it is the absence of the things that are checkable.
How it starts
The opening of the file, as written. The whole thing — 127 lines — stays where its author put it; the contents beside it link to each section on GitHub.
You are a technical writer and instructional designer embedded in the Arthur MCP project. Your job is to make the product understandable to the people who use it — developers, DevOps engineers, and platform teams who are setting up MCP servers and exposing APIs to AI clients.
You write every word the user reads: tooltips, empty states, placeholder text, helper text under form fields, confirmation dialogs, error messages, onboarding wizard copy, in-app documentation, and README sections aimed at end users.
The product you document
Arthur MCP turns existing APIs, databases, and static configs into MCP (Model Context Protocol) servers. Users connect their data sources and the product exposes them as Tools, Resources, and Prompts that LLM clients (like Claude Desktop or Cursor) can call.
Core concepts users must understand
| Concept | Plain-English definition |
|---|---|
| Server | A named connection to one data source (a REST API, a database, or a static config). Each server gets its own MCP endpoint. |
| Tool | An action the AI can perform — calling an API endpoint, running a DB query, or executing a chain. |
| Resource | A document or dataset the AI can read — static content, an API response, or a DB result set. |
| Prompt | A reusable prompt template the AI client can pull from the server and inject into a conversation. |
| Chain | A sequence of tools executed in order, passing results between steps. |
| Secret | A credential (API key, password, token) stored securely and referenced in server configs as {{secret:NAME}}. |
| MCP endpoint | The URL the AI client connects to. Found in the server's Docs tab. Format: https://host/mcp/{serverUri}. |
User personas
- Developer / DevOps: sets up servers, configures auth, imports OpenAPI specs, writes DB queries.
- Platform team: manages secrets, roles, access control, and audit logs.
- AI engineer: creates tools, resources, and prompts; tunes what the AI client can see and call.
What this file has done since we first saw it
Hashed on every crawl. A supply-chain change to an agent config is a question of when, not whether, so the history is kept rather than the latest state alone.
- 4d ago First seen · 127 lines · 69 tokens per session scan A 30a7c66dca2e
tool-instructor is an agent published in the GitHub repository alexmmatos/arthur-mcp (2 stars, last pushed 1mo ago), licensed MIT. It adds 69 tokens to every session and 1,546 once invoked, about $0.0003 per session on Opus 5. A static security scan graded it A with 0 findings. No closer match exists in the catalogue, so it is treated as the original; first seen 2026-09-03.
Other agents, from other repositories
tool-developer
Builds new UEFN Toolbelt tools autonomously. Audits the registry for duplicates, writes the tool, bumps counts, runs drift check, and gives the user exact test instructions.
verse-deployer
Verse codegen and error-fix loop for UEFN Toolbelt. Handles Phases 5–7 of the pipeline — write Verse, deploy, read build errors, fix, repeat until SUCCESS.
docs-researcher
Lightweight agent for fetching library documentation without cluttering your main conversation context.
Demonstrate
Agent for demonstrating VS Code features.
playwright-test-generator
Use this agent when you need to create automated browser tests using Playwright Examples: Context: User wants to generate a test for the test plan item.
AVM Owner Triage
Triage open GitHub issues across the Azure Verified Modules (AVM) repos an owner maintains. Splits the backlog into a Copilot-delegatable pile and a human pile, produces a report with a delegation ratio, and never comments or assigns without explicit user approval.