MCP Hub
Back to servers

Code Whisperer MCP

Provides access to a library of 18 specialized skills, project templates, and prompt patterns for Claude Code. It enables automated workflows for feature development, security audits, and system maintenance through multi-agent collaboration.

glama
Updated
Mar 30, 2026

@code-whisperer/skills

MCP server for The Code Whisperer — 18 battle-tested Claude Code skills, 5 CLAUDE.md templates, and 10+ prompt patterns from a system running 49+ apps, 6 live bots, and real money on the line.

Install once. Use everywhere.

Prerequisites

  • Node.js >= 18
  • Claude Code CLI or any MCP-compatible client (Cursor, Windsurf)
  • A Code Whisperer API key — get one at jeremyknox.ai/skills-library

Install

Add to your ~/.claude.json (Claude Code) or equivalent MCP config:

{
  "mcpServers": {
    "code-whisperer": {
      "command": "npx",
      "args": ["-y", "@code-whisperer/skills@latest"],
      "env": {
        "CODEWHISPERER_API_KEY": "<your-key>"
      }
    }
  }
}

Restart Claude Code. The tools are now available in every session.

Tools

ToolDescription
list_skillsList all 18 skills with slug, category, and description
get_skill(slug)Get the full SKILL.md content for a skill
list_templatesList all 5 CLAUDE.md templates
get_template(slug)Get the full CLAUDE.md template content
list_promptsList all prompt patterns with tags
get_prompt(slug)Get the full prompt content

Usage in Claude Code

Once installed, ask Claude to use the tools directly:

Use the code-whisperer MCP to get the feature-team skill

Or reference skills by name and Claude will fetch them:

/feature-team build the new dashboard page per the PRD

Available Skills

SlugCategoryWhat it does
feature-teamAgent Teams3-agent (Backend + Frontend + QA) feature implementation
quality-teamAgent TeamsPre-release QA gate + coverage remediation
security-teamAgent TeamsSecurity review (Static + Dependency + Threat Model)
design-teamAgent TeamsUI implementation + accessibility + design system
research-teamAgent TeamsMarket research, competitive intel, pre-PRD discovery
audit-swarmAudit & ReviewParallel multi-agent code audit with P0/P1 tickets
audit-swarm-resolveAudit & ReviewResolve audit findings with fix agents per project
repo-maintenanceAudit & ReviewCLAUDE.md pruning, stub test detection, CI optimization
incidentDevOps & OpsP0 incident response: halt, ticket, logs, timeline, postmortem
onboard-repoDevOps & OpsFull repo onboarding: sec-scan, CLAUDE.md, CI, quality gates
debug-investigateIntelligenceScientific method debugging with hypothesis elimination log
deep-diveIntelligenceExtract knowledge from YouTube, URLs, or topics → integration plan
document-swarmContent & RetroGenerate READMEs, runbooks, API docs via agent team
morning-briefContent & RetroExecutive morning briefing: overnight status + priorities
weekly-retroContent & RetroWeekly 'State of the Empire' brief stored to knowledge base
trade-retroContent & RetroP&L review and trading pattern analysis
skillbossBuild & DeployMulti-AI gateway: Cloudflare Workers, D1/KV, Stripe, auth, AI
academy-trackBuild & DeployGenerate full Academy tracks with lessons, quizzes, and CI
blog-autopilotContent & RetroAutomated content pipeline: YouTube → article → PR

Available CLAUDE.md Templates

SlugWhat it covers
nextjs-saasNext.js 14 App Router SaaS — auth, DB, Stripe, Vercel
python-servicePython microservice — FastAPI, launchd, Discord alerts
discord-botDiscord bot with slash commands and D1 persistence
react-viteReact + Vite component library or standalone app
python-cliPython CLI / automation script — argparse, logging, CI

Authentication

Pass your API key as the api_key argument to any tool, or set CODEWHISPERER_API_KEY in the MCP server env (recommended — set once in your config, never passed again).

If CODEWHISPERER_API_KEY is not set in the server environment, the server runs in unauthenticated mode (dev/testing only).

License

MIT — Jeremy Knox

Reviews

No reviews yet

Sign in to write a review