Mcp Prompt Library — independently scanned and version-tracked by SaferSkills.
SaferSkills independently audited Mcp Prompt Library (Agent Skill) and scored it 65/100 (yellow). The audit ran 55 deterministic rules across Security, Supply Chain, Maintenance, Transparency, and Community; it found 5 high-severity and 4 lower-severity findings. The full rule-by-rule trace and per-finding evidence are below. Free, methodology-open.
Findings & checks · 9 flagged
The text {match} is the classic direct prompt-injection phrasing. Placed in a skill body that the agent reads as trusted instructions, it tries to make the agent abandon its prior rules and follow whatever comes next — a full system-prompt override.
ignore/disregard/forget … previous instructions sentence.The text {match} is the classic direct prompt-injection phrasing. Placed in a skill body that the agent reads as trusted instructions, it tries to make the agent abandon its prior rules and follow whatever comes next — a full system-prompt override.
ignore/disregard/forget … previous instructions sentence.The text {match} tells the agent to skip the normal "ask the user first" gate. Used adversarially it removes the human-in-the-loop check before destructive or sensitive actions, turning a normally-gated agent into a fire-and-forget executor.
A bulleted imperative like {match} tells the agent to never reveal, disclose, or mention something to the user. Used adversarially it can instruct the agent to hide its tool calls or lie about what it did — stripping the transparency a user relies on to trust the agent.
A bulleted imperative like {match} tells the agent to never reveal, disclose, or mention something to the user. Used adversarially it can instruct the agent to hide its tool calls or lie about what it did — stripping the transparency a user relies on to trust the agent.
Every scanned point with the score it earned and what moved between them.
First recorded scan — no prior version to compare against.
The primary manifest — the file an agent reads to learn what this artifact does.
<p align="center"> <h1 align="center">MCP Prompt Library</h1> <p align="center"> <strong>110+ curated prompts, workflows, and coding standards for AI-assisted development</strong> </p> </p>
<p align="center"> <a href="LICENSE"><img src="https://img.shields.io/badge/license-MIT-blue.svg" alt="License"></a> <a href="https://nodejs.org"><img src="https://img.shields.io/badge/node-%3E%3D18.0.0-brightgreen.svg" alt="Node.js"></a> <a href="https://modelcontextprotocol.io"><img src="https://img.shields.io/badge/MCP-compatible-purple.svg" alt="MCP Compatible"></a> <img src="https://img.shields.io/badge/tests-147%20passed-success.svg" alt="Tests"> </p>
<p align="center"> <a href="#quick-start">Quick Start</a> • <a href="#features">Features</a> • <a href="#mcp-server">MCP Server</a> • <a href="#library-contents">Library</a> • <a href="#integrations">Integrations</a> </p>
AI assistants are powerful, but they're only as good as the prompts you give them. Most developers:
MCP Prompt Library solves this by providing:
| What | How |
|---|---|
| 110+ battle-tested prompts | Organized by development phase (planning, development, quality, operations, performance, maintenance, devex, security, evals) |
| MCP server with 15 tools | Direct integration with Claude, OpenCode, Cursor, and any MCP-compatible client |
| Smart suggestions | AI recommends prompts based on what you're doing |
| Workflow chains | Multi-step guided processes for features, bugs, refactoring, security |
| Composable snippets | Mix modifiers like ultrathink + security-first on any prompt |
Install the library to use prompts programmatically in your own tools:
npm install @esreekarreddy/ai-promptsimport {
getPrompt,
searchPrompts,
composePrompts,
} from "@esreekarreddy/ai-prompts";
// 1. Get a specific prompt
const prd = getPrompt("prd-generator");
console.log(prd.content);
// 2. Search for prompts
const securityPrompts = searchPrompts("security");
// 3. Compose a custom prompt with modifiers
const complexTask = composePrompts([
"prd-generator",
"ultrathink",
"security-first",
]);git clone https://github.com/esreekarreddy/mcp-prompt-library.git
cd mcp-prompt-library/mcp-server
npm install && npm run buildSee Integrations for your specific tool (OpenCode, Claude Desktop, Cursor).
Once connected, your AI assistant has access to all prompts:
"suggest prompts for what I'm doing"
"get the PRD generator prompt"
"start the new-feature chain"
"compose prd-generator with ultrathink"Tell your AI what you're doing, and it suggests the right prompts:
| You Say | It Suggests |
|---|---|
| "I need to build a new feature" | PRD generator, new-feature chain |
| "Stuck on a bug" | Deep debugger, debugging skill, bug-fix chain |
| "Security review before launch" | Security audit, security-hardening chain |
| "This code is a mess" | Code cleaner, refactoring skill, refactor chain |
| "Complex architecture decision" | Megathink modifier, senior-engineer persona |
Multi-step guided workflows for complex tasks:
| Chain | Steps | Use Case |
|---|---|---|
new-feature | 7 | From PRD to deployment |
bug-fix | 7 | Systematic debugging to resolution |
refactor | 6 | Safe refactoring with verification |
security-hardening | 7 | Comprehensive security review |
production-launch | 7 | Pre-launch checklist to deployment |
# Start a chain
node dist/cli.js chains # List available chains
start_chain chain="new-feature" # Via MCP toolCombine any prompts with modifiers:
# CLI
node dist/cli.js compose prd-generator ultrathink security-first
# MCP Tool
compose_prompt items=["prd-generator", "ultrathink", "security-first"]Instant prompt enhancers:
| Modifier | Effect |
|---|---|
ultrathink | Deep analysis with extended reasoning |
megathink | Maximum thinking for architecture decisions |
critique | Harsh, unfiltered feedback mode |
debug | Systematic debugging approach |
plan | Planning mode - no code yet |
secure | Security-focused review |
simplify | Explain like I'm 12 |
The MCP (Model Context Protocol) server exposes 15 tools to your AI assistant:
| Tool | Purpose |
|---|---|
get_prompt | Fetch any prompt by name (fuzzy matching works) |
search_prompts | Search library by keywords |
suggest_prompts | Smart suggestions based on your intent |
enhance_prompt | Analyze request and suggest approach + relevant prompts |
save_to_library | Save new prompts to the library |
library_stats | Library statistics |
random_prompt | Random prompt for inspiration |
| Tool | Purpose |
|---|---|
list_chains | View available workflow chains |
start_chain | Begin a multi-step workflow |
chain_next | Advance to next step |
chain_status | View workflow progress |
chain_step | Jump to specific step |
| Tool | Purpose |
|---|---|
compose_prompt | Combine multiple prompts |
quick_prompt | Instant one-liner modifiers |
detect_context | Analyze project → suggest stack-specific prompts |
You: "Build a user authentication system"
↓
AI calls: suggest_prompts("Build a user authentication system")
↓
Returns: security-audit, new-feature chain, auth patterns
↓
AI calls: start_chain("new-feature")
↓
AI guides you through: PRD → Architecture → Implementation → Testing → Deploy110+ curated resources across 8 categories:
mcp-prompt-library/
├── prompts/ (38) - Copy-paste ready prompts
│ ├── planning/ PRD generator, scope killer, architecture
│ ├── development/ Debugger, code cleaner, tech debt
│ ├── quality/ Security audit, PR reviewer, changelog generator
│ ├── operations/ Observability, runbooks, incident helper, postmortem
│ ├── performance/ Profiler, load test planner, optimization guide
│ ├── maintenance/ Dependency upgrader, migration planner, tech spec
│ ├── devex/ Repo onboarding, CI fixer, codebase explainer
│ ├── security/ Prompt injection audit
│ ├── evals/ Response grader, consistency checker, quality rubric
│ ├── design/ Design system extractor
│ ├── analysis/ Deep debugger
│ ├── agentic/ Context manager, agentic loop, test-driven fix
│ └── system/ Master system prompt for AI setup
├── skills/ (8) - AI behavior definitions
│ └── code-review, debugging, testing, refactoring, documentation...
├── instructions/ (18) - Reusable system prompts
│ ├── personas/ Senior engineer, security expert, DevOps, UX
│ ├── standards/ TypeScript, React, Python, Go, Rust, FastAPI, Next.js
│ └── workflows/ TDD, PR review, incident response, feature development
├── templates/ (16) - Project scaffolding
│ ├── claude-md/ CLAUDE.md for Next.js, Python, Node.js, CLI tools
│ ├── cursor-rules/ .cursorrules for various stacks
│ ├── copilot/ GitHub Copilot instructions
│ └── docs/ PRD, ADR, API spec, runbook templates
├── chains/ (5) - Multi-step workflows
│ └── new-feature, bug-fix, refactor, security-hardening, production-launch
├── snippets/ (21) - Composable modifiers
│ ├── modifiers/ ultrathink, megathink, step-by-step, meta-cot
│ ├── output-formats/ JSON, markdown table, checklist, numbered list
│ ├── constraints/ Security first, MVP only, read-only, no external deps
│ └── safety/ Fact-check, citation-required, uncertainty, injection-guard
├── contexts/ (9) - Reference documentation
│ ├── stacks/ Next.js 14, FastAPI, Prisma
│ ├── patterns/ MCP server patterns, agentic coding
│ └── guides/ API design, error handling
└── examples/ (3) - Gold-standard samples
└── PRDs, architecture docs, code reviewsAdd to ~/.opencode/config.json:
{
"mcp": {
"ai-library": {
"type": "local",
"command": [
"node",
"/path/to/mcp-prompt-library/mcp-server/dist/index.js"
],
"enabled": true
}
}
}Add to ~/Library/Application Support/Claude/claude_desktop_config.json:
{
"mcpServers": {
"ai-library": {
"command": "node",
"args": ["/path/to/mcp-prompt-library/mcp-server/dist/index.js"]
}
}
}Add to .cursor/mcp.json in your project:
{
"mcpServers": {
"ai-library": {
"command": "node",
"args": ["/path/to/mcp-prompt-library/mcp-server/dist/index.js"]
}
}
}Copy the prompt files to your workspace:
cp -r .github/prompts /path/to/your-project/.github/promptsUse the library directly from your terminal:
cd mcp-server
# Get a specific prompt
node dist/cli.js get prd-generator
# Search prompts
node dist/cli.js search "security"
# Get AI-powered suggestions
node dist/cli.js suggest "I need to refactor this messy code"
# Combine prompts
node dist/cli.js compose prd-generator ultrathink step-by-step
# View workflow chains
node dist/cli.js chains
# Library statistics
node dist/cli.js stats
# Random prompt for inspiration
node dist/cli.js randomcd mcp-server
npm install # Install dependencies
npm run build # Build TypeScript
npm run test # Run tests (114 tests)
npm run dev # Watch mode
npm run typecheck # Type checking onlySee CONTRIBUTING.md for guidelines.
Ideas for contributions:
MIT - see LICENSE
Built with the Model Context Protocol by Anthropic.
<p align="center"> <strong>Your AI is only as good as your prompts. Keep them polished.</strong> </p>
~30 seconds. Free. No account. Every finding cites a rule and a line of evidence.