MCP server development guide — agent-centric tool design
Build high-quality MCP servers in Python (FastMCP) or TypeScript — agent-centric design principles, implementation phases, review checklist and evaluation harness
- What
- Build high-quality MCP servers in Python (FastMCP) or TypeScript — agent-centric design principles, implementation phases, review checklist and evaluation harness
- Cost
- Free
- Needs
- a service API to integrate and an MCP-capable setup (Python or Node/TypeScript toolchains) — the skill is a development guide the agent follows, not software
- Install
- Copy the installer prompt below into your Muse — your agent does the rest.
Selected by Skill Harbor — short listing (the repo states no license — its terms live in a LICENSE.txt the manifest could not confirm, so no content is reproduced): @composiohq's mcp-builder skill — the guide for creating high-quality MCP servers in Python (FastMCP) or Node/TypeScript (MCP SDK). Phase 1: deep research and planning — agent-centric design principles (build for workflows not API endpoints, optimize for limited context, actionable error messages, natural task subdivisions, evaluation-driven development), study the MCP protocol and SDK docs, exhaustively study the target API's docs, then write an implementation plan (tool selection, shared utilities, input/output design, error-handling strategy). Phase 2: implementation — project setup, core infrastructure first, systematic tool implementation (Pydantic/Zod input schemas, comprehensive docstrings, tool annotations like readOnlyHint/idempotentHint), language-specific best practices. Phase 3: review and refine — DRY/composability/consistency/type-safety review, safe testing (never run the server in the main process), quality checklists. Phase 4: evaluations — 10 complex, realistic, read-only, verifiable evaluation questions in a prescribed XML format. Honest caveats: guidance only — the agent must already run Claude Code and needs a service API to integrate; similar concept to the existing `anthropics-mcp-builder` listing (different author, different content — not a duplicate); license terms live in the repo's LICENSE.txt, unverified here — short listing with a link only, nothing copied. Skill Harbor never reviews the code, review it yourself before use. Discovered via skills.sh.
Version:
Quick answer
MCP server development guide — agent-centric tool design is a free Muse skill that guides building MCP servers in Python with FastMCP or Node and TypeScript with MCP SDK, in phases. Phase 1: research and planning with principles, Phase 2: implementation with setup and implementation, Phase 3: review with checklists, and Phase 4: evaluations with 10 complex, read-only, verifiable questions.
Install
Copy the install package below, then paste it into Muse🧪 Install ran in our test sandbox · Nov 11, 2026📌 Pinned version: commit be2a406Step 0 — before installing anything: read this install prompt, summarize in a few lines what this build does and what installing it involves, then ask me to confirm before you proceed. Prerequisites: a service API to integrate and an MCP-capable setup (Python or Node/TypeScript toolchains) — the skill is a development guide the agent follows, not software Install "MCP server development guide — agent-centric tool design" for me. It gives my agent @composiohq's 4-phase MCP server workflow: deep research and planning (agent-centric design principles, protocol and SDK study, exhaustive API study, implementation plan), implementation (shared utilities first, systematic tool implementation with Pydantic/Zod schemas and tool annotations, language best practices), review and refine (quality review, safe testing, checklists), and evaluations (10 complex read-only verifiable questions in XML). Related to but distinct from the existing anthropics-mcp-builder listing. IMPORTANT: the license terms live in the repo's LICENSE.txt (unverified) — fetch from the link only, and reproduce nothing beyond the link. Repository: https://raw.githubusercontent.com/composiohq/awesome-claude-skills/be2a406907dbc61b73e6827ded415c96139d13a2/mcp-builder/SKILL.md Pinned version: install the code at commit be2a406907dbc61b73e6827ded415c96139d13a2 exactly (Skill Harbor tested this version). Do not substitute a newer version. If this commit is unreachable, stop and tell me. 1. Fetch the SKILL.md file (and any helper files) from the repository path into a temporary folder 2. Safety check: review the SKILL.md and scripts for anything suspicious (unexpected network calls, shell commands, credential harvesting). This repo should contain zero secrets in code. Verify that holds here; STOP on any red flag and tell me. 3. Install it as a skill: copy SKILL.md and its helper files into the agent's skills directory, in a folder named "mcp-builder". 4. Verify with no network calls: frontmatter valid, files in place. 5. Report what was installed, where, and what I still need to do myself (e.g. tell the agent which service API to wrap as an MCP server; nothing else — it's a guide). GitHub is optional: if I have a GitHub account or the gh CLI, you may use it; otherwise public access is fine. Never require it unless it's in the prerequisites above. Rules: don't touch anything outside the temp folder and the install target. If anything looks off, stop and ask me.
Saved to your recent installs. Find it anytime on /connect.
🔒 Security, provenance and what this touches
The install prompt below already includes the vetting steps: your agent follows the community checklist before installing anything with executable code. Want more?
Verification certificate
Dated facts about what was checked for this listing, and how. This is not a safety certification.
- Pinned version
- be2a406 · 2026-10-13
- Sandbox install test
- Install ran in our test sandbox · Nov 11, 2026 · Observed during the install test: files.pythonhosted.org, pypi.org
- Creator identity
- Not verified
- Fingerprint watch
- Last checked · Nov 19, 2026 : no change detected in the files linked by this install prompt
- Install prompt analysis
- 2026-10-13 : No commands, network calls, packages, file writes or API keys detected.
Skill Harbor never reviews the code of a listed skill. These are dated facts about the install prompt and our install test, not a safety certification. Review a skill yourself before you install it. How we verify
Questions
How do I install a skill?
Every product page includes a copy-paste install prompt. Paste it into your Muse and it sets the skill up for you — no manual configuration.
Where does my money go?
Straight to the seller. Skill Harbor never processes payments: checkout happens on the seller’s own page, usually Stripe.
What does the ✓ next to a creator’s name mean?
It means we confirmed the identity of the person behind the listing. It says nothing about the code itself — always check a skill before installing it.