Airis-mcp-gateway mcp-implementation

Use when implementing features that use external libraries or APIs — ensures docs-first approach before writing code

install
source · Clone the upstream repo
git clone https://github.com/agiletec-inc/airis-mcp-gateway
Claude Code · Install into ~/.claude/skills/
T=$(mktemp -d) && git clone --depth=1 https://github.com/agiletec-inc/airis-mcp-gateway "$T" && mkdir -p ~/.claude/skills && cp -r "$T/skills/skills/mcp-implementation" ~/.claude/skills/agiletec-inc-airis-mcp-gateway-mcp-implementation && rm -rf "$T"
manifest: skills/skills/mcp-implementation/SKILL.md
source content

MCP Docs-First Implementation Workflow

Use this workflow when implementing features that involve external libraries or APIs you're not 100% certain about.

Prerequisites

Gateway instructions handle tool routing. This skill teaches the implementation sequence that avoids wasted effort from wrong assumptions.

Workflow

Step 1: Check documentation before coding

Before writing any implementation code:

  1. Look up the library/API documentation using Gateway doc tools
  2. Find the specific function/method/endpoint you plan to use
  3. Check for official examples or sample code

If official examples exist, use them as your starting point rather than writing from scratch.

Step 2: Check existing patterns in the codebase

Use native Grep/Glob tools (not Gateway) to:

  1. Search for existing usage of the same library in this project
  2. Check import patterns and initialization conventions
  3. Look for existing wrappers or utilities

If the codebase already uses this library, follow the established pattern.

Step 3: Implement

With documentation and existing patterns in hand:

  1. Write the implementation following official docs and project conventions
  2. Handle errors explicitly — no silent fallbacks
  3. Keep the implementation minimal — don't add features beyond what's needed

Step 4: Verify

  1. Run tests to confirm the implementation works
  2. Check for type errors and lint issues
  3. If the change is user-facing, verify in the browser

Decision Points

SituationAction
Using a library for the first timeFull workflow: docs → patterns → implement → verify
Library already used in codebaseSkip docs, follow existing patterns
API call to external serviceAlways check docs for auth, rate limits, error codes
Uncertain about behaviorCheck docs first, don't guess

Anti-patterns

  • Writing code based on memory instead of checking current docs
  • Ignoring existing patterns in the codebase
  • Adding error handling for scenarios that can't happen
  • Over-engineering beyond the task requirements