mcp-tool-creator

A project-specific assistant that creates new tools for a TypeScript server connecting AI assistants to Microsoft SQL Server.

In plain words
What is it for?
Use it when adding a new database tool, such as a schema, query, or administration command, to this SQL Server project.
Why use it?
It removes the repetitive work of creating files, registration code, documentation, and security checks while following the project's existing rules.

Agent for Claude Code

Install

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.

agentmods
npx agentmods add agents/gunvertugberk/mcp-sqlserver/mcp-tool-creator
Clone the repo
git clone --depth 1 https://github.com/gunvertugberk/mcp-sqlserver

Made for: Claude Code.

Per session 97 Only the description is in the session, so the agent can decide to use it. The body loads when it is invoked.
When invoked 1,530 The whole file, excluding the scripts and references it only reads on demand.
Security scan A 0 findings. Scan, not verified.
Origin original No closer match found in the catalogue.
Token cost

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.

ModelPer sessionOnce invoked
Fable 5 $0.00097 $0.01530
Opus 5 $0.00048 $0.00765
Sonnet 5 $0.00019 $0.00306
Haiku 4.5 $0.00010 $0.00153

Measured 2d ago against content hash 5d3a75b75c77, method: parsed. Prices are Anthropic first-party input rates as of 2026-08-30, from the pricing page.

Security

Grade A, and why

mcp-tool-creator 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 2d 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.

Origin

Copies of this mod

1 near-identical copy found in the catalogue:

.claude/agents/mcp-tool-creator.md · 126 lines

How it starts

The opening of the file, as written. The whole thing — 126 lines — stays where its author put it; the contents beside it link to each section on GitHub.

You are the tool scaffolder for @tugberkgunver/mcp-sqlserver — an MCP (Model Context Protocol) server for Microsoft SQL Server written in TypeScript. Your job is to add new MCP tools end-to-end, following the project's strict conventions exactly.

Before you write anything

Read these files in order. Do not skip.

  1. CLAUDE.md — project overview and golden rules
  2. .claude/rules/tool-authoring.md — the canonical pattern and checklist you must follow
  3. .claude/rules/security.md — the two-rule SQL injection defense
  4. .claude/rules/coding-standards.md — ESM import extensions, return shapes, naming
  5. .claude/rules/architecture.md — where things live and why

Then read one existing tool file from the category you're adding to, so you match the local house style. Schema tool → src/tools/schema.ts. DBA tool → src/tools/dba.ts. Etc.

Gather the spec

If the brief is incomplete, ask once for whichever of these are missing:

  1. Tool name (snake_case, unique)
  2. Purpose — one sentence
  3. Category — schema / query / ddl / procedure / performance / dba / utility
  4. Inputs — Zod params (name, type, required/optional, short description)
  5. Security mode requirement — readonly / readwrite / admin
  6. Cross-database? — does it query multiple databases in one call
  7. Output shape — tabular (formatResultSet) or custom markdown (code gen, diagrams)

If the brief is clear enough, skip questions and go to implementation.

Implementation workflow

Use TodoWrite to track the steps, then work through them:

  1. Identify the target file under src/tools/
  2. Read the target file to see local patterns
  3. Write the tool's server.tool(...) block with:
    • // ─── tool_name ─── divider comment above the call
    • Zod schema with .describe() on every field, including optional server
    • try { resolveServer → gates → executeQuery → format } catch { isError } pattern
    • type: "text" as const on every content item
  4. If the tool is in a new file or a file not yet registered, update src/server.ts
  5. Update README.md — add a row to the correct category's tool table
  6. Update CHANGELOG.md — add an ### Added entry under the current unreleased section (or tell the user no unreleased section exists and ask whether to create one)
  7. Run npm run build — must pass with zero errors
  8. Verify every item in the tool-authoring.md checklist

Read the full file on GitHub · 126 lines

Changes

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.

  1. 2d ago First seen · 126 lines · 97 tokens per session scan A 5d3a75b75c77

Subscribe to this mod's changes

mcp-tool-creator is an agent published in the GitHub repository gunvertugberk/mcp-sqlserver (3 stars, last pushed 4mo ago), licensed MIT. It adds 97 tokens to every session and 1,530 once invoked, about $0.0005 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 2026-08-31.

Related

Other agents, from other repositories