Skip to content

seb1n/awesome-ai-agent-skills / agent-engineering/mcp-server-building

mcp-server-building

Verified

Design, implement, harden, and verify Model Context Protocol (MCP) servers with precise tool contracts, least-privilege authorization, safe transports, structured errors, and interoperability tests. Use when creating a new MCP server, exposing an API or data source through MCP, reviewing an MCP server design, adding or revising MCP tools, or preparing an MCP server for production.

FileKindSize
SKILL.mdskill md9.7 KB
agents/openai.yamlconfig228 B
assets/server-design-template.mdmarkdown1.8 KB
references/server-design-checklist.mdmarkdown6.9 KB
scripts/validate_tool_manifest.pyexecutablescript11.5 KB

assets/server-design-template.md

View on GitHub
# MCP server design

## Decision summary

- User job:
- In scope:
- Out of scope:
- Supported clients:
- Transport:
- Protocol revision(s):
- Runtime and SDK:
- Owner:

## Trust and data flow

Describe host, client, server, downstream systems, identities, credentials, trust boundaries, and sensitive data.

## Tool catalog and manifest

| Tool | Purpose | Effects | Data classes | Authorization mode / enforcement | Idempotency | Input / output / error schemas |
|---|---|---|---|---|---|---|
|  |  |  |  |  |  |  |

Represent effects and data classes as arrays, even when each contains one value. Give every tool an authorization object; use `mode: public` explicitly rather than omitting authorization. Store machine-checkable `inputSchema`, `outputSchema`, and `errorSchemas` in the JSON manifest. A passing structural lint is not protocol conformance or a safety certification.

## Authorization model

- Authentication:
- Tenant/object/action checks:
- Scopes:
- Downstream token exchange:
- Approval requirements:
- Credential storage and rotation:

## Reliability and limits

- Timeouts:
- Retry policy:
- Concurrency and rate limits:
- Pagination and payload limits:
- Cancellation:

## Verification plan

| Requirement | Test | Expected result | Evidence |
|---|---|---|---|
|  |  |  |  |

For revision `2026-07-28`, include `server/discover`, per-request `_meta`, `Mcp-Method`/`Mcp-Name` header agreement on Streamable HTTP, required `resultType`, cache metadata, and MRTR where applicable. Do not test `initialize` or `Mcp-Session-Id` as the current path; test them only for explicitly supported legacy revisions.

## Operations

- Logs and audit events:
- Metrics and alerts:
- Deployment strategy:
- Rollback:
- Credential revocation:

## Assumptions and residual risk

- Assumptions:
- Unverified items:
- Residual risks and owners:
finishing-a-development-branchobra/superpowersUse when implementation is complete, all tests pass, and you need to decide how to integrate the workStars: 322kUpdated Claude CodeCodexVerifiedInstall
systematic-debuggingobra/superpowersUse when encountering any bug, test failure, or unexpected behavior, before proposing fixesStars: 322kUpdated Claude CodeCodexVerifiedInstall
test-driven-developmentobra/superpowersUse when implementing any feature or bugfix, before writing implementation codeStars: 322kUpdated Claude CodeCodexVerifiedInstall
skill-creatoranthropics/skillsCreate new skills, modify and improve existing skills, and measure skill performance. Use when users want to create a skill from scratch, edit, or optimize an existing skill, run evals to test a skill, benchmark skill performance with variance analysis, or optimize a skill's description for better triggering accuracy.Stars: 198kUpdated Claude CodeCodexVerifiedInstall