⚙️ PromptOS

Enterprise AI Engineering Control Plane
Govern, version, test, and deploy prompts like code.

🧭 What is PromptOS?

PromptOS is a full-stack AI prompt engineering platform built by JourdanLabs × C&L Strategy. It treats prompts the way modern engineering treats code — with versioning, testing, access control, deployment pipelines, and observability.

Whether you're a solo developer or a team of 50 engineers, PromptOS gives you:

🌐 Web App — A beautiful UI with 3 themes, 9 prompt templates, and 5 LLM presets
🖥️ CLI — Full prompt execution, registry management, analytics, and pack tooling from the terminal
🧩 IDE Extension — Run prompts directly in VSCode or Cursor without leaving your editor
🏗️ Platform — Runtime, RBAC, Policy Engine, CI Eval, Semantic Diff, Analytics, and a Learning Loop

🚀 Quick Start

# ── Web App ────────────────────────────────────────
git clone https://github.com/lelandsequel/prompt-library
cd prompt-library
npm install
npm run dev
# → Open http://localhost:3000

# ── CLI ────────────────────────────────────────────
node promptos/cli/index.js list
node promptos/cli/index.js run python-function --dry-run

# ── IDE Extension ──────────────────────────────────
code --install-extension promptos/ide/extension/promptos-1.0.0.vsix

# ── Registry Server ────────────────────────────────
cd promptos/server && npm install && node server.js --port 3001

🌐 Web App (`/src/app/`)

A Next.js 15 application with three switchable visual themes, an interactive prompt builder, and real-time preview.

🎨 Themes

Theme	Description
Brutalist	Bold, high-contrast, monochrome — built for focus
Minimal	Clean, spacious, elegant — the default experience
Terminal	Dark, file-tree sidebar, hacker aesthetic

Switch themes at any time using the toggle in the header. Your selection persists across the session.

📋 Prompt Templates (9 built-in)

Template	Category	Description
Code Review	Engineering	Deep review for quality, bugs, and best practices
Debugging	Engineering	Systematic root-cause analysis
Architecture Design	Architecture	System design with trade-offs and diagrams
API Design	Engineering	REST/GraphQL API spec generation
Incident Response	SRE	Structured incident runbook creation
Test Cases	QA	Comprehensive test suite generation
Refactoring	Engineering	Safe, structured code improvement
Security Review	Security	Threat modeling and vulnerability analysis
Performance Optimization	Engineering	Profiling-guided performance improvements

🤖 LLM Presets (5)

Each preset formats the generated prompt optimally for the target model:

Preset	Optimized For
ChatGPT	OpenAI GPT-4 / GPT-4o
Claude	Anthropic Claude Sonnet / Opus
Cursor	Cursor AI IDE assistant
Copilot	GitHub Copilot Chat
Gemini	Google Gemini Pro / Ultra

✨ Features

Category filter — Browse by Engineering, Architecture, SRE, Security, QA
Search — Live search across template names and descriptions
Variable fill-in — Required fields (*) and optional fields for richer output
Real-time preview — Generated prompt updates as you fill fields
One-click copy — Copy to clipboard instantly
Custom Prompt Optimizer — Paste any prompt and AI-optimize it for your target LLM

🖥️ CLI (`promptos/cli/index.js`)

The PromptOS CLI is a full-featured terminal tool for managing and executing prompts.

# ── Listing & Discovery ────────────────────────────
node promptos/cli/index.js list                   # List all prompts
node promptos/cli/index.js list --tag codegen     # Filter by tag

# ── Running Prompts ────────────────────────────────
node promptos/cli/index.js run <prompt_id>        # Run with defaults
node promptos/cli/index.js run <prompt_id> \
  --model claude-sonnet \
  --user alice \
  --role engineer                                 # Run with enforcement
node promptos/cli/index.js run <prompt_id> \
  --dry-run                                       # Render only, no API call

# ── Validation ─────────────────────────────────────
node promptos/cli/index.js validate               # Validate all prompts
node promptos/cli/index.js validate <prompt_id>   # Validate one prompt

# ── Analytics ──────────────────────────────────────
node promptos/cli/index.js analytics report       # Summary report
node promptos/cli/index.js analytics export \
  --output usage.json                             # Export telemetry

# ── Packs ──────────────────────────────────────────
node promptos/cli/index.js pack list              # List installed packs
node promptos/cli/index.js pack install <pack.yaml>
node promptos/cli/index.js pack uninstall <pack-name>
node promptos/cli/index.js pack build <dir>       # Build a pack from dir

# ── Semantic Diff ──────────────────────────────────
node promptos/cli/index.js diff v1.yaml v2.yaml   # Compare prompt versions

# ── Registry Server ────────────────────────────────
node promptos/cli/index.js server --port 3001     # Start local server

🏗️ PromptOS Platform (`/promptos/`)

📚 Prompt Registry (`promptos/registry/`)

The registry is the source of truth for all prompts in your organization.

promptos/registry/
├── index.yaml          ← Human-readable registry index
├── index.json          ← Machine-readable registry (served by API)

11 built-in prompts as PromptSpec YAML files, organized by domain:

promptos/prompts/
├── architecture/
│   └── architecture-v1.yaml
├── codegen/
│   ├── api-design-v1.yaml
│   ├── code-review-v1.yaml
│   ├── codegen-v1.yaml
│   ├── performance-v1.yaml
│   ├── refactoring-v1.yaml
│   ├── security-review-v1.yaml
│   └── test-cases-v1.yaml
├── debugging/
│   └── debugging-v1.yaml
├── docs/
│   └── docs-gen-v1.yaml
└── sre/
    └── incident-response-v1.yaml

PromptSpec Schema:

id: code-review
version: 1.0.0
role: engineer
description: "Deep code review for quality, correctness, and style"
inputs:
  schema:
    language: { type: string, required: true }
    code: { type: string, required: true }
outputs:
  schema:
    review: { type: string }
prompt: |
  You are an expert {{language}} engineer.
  Review the following code for quality, bugs, and best practices:
  {{code}}

⚡ Runtime (`promptos/runtime/` + `promptos/execute.js`)

Every prompt execution flows through a unified, observable pipeline:

execute(prompt_id, inputs, options)
        │
        ▼
  1. RBAC Check       ← Does this user/role have run_prompt permission?
        │
        ▼
  2. Policy Check     ← Is this model allowed? Does data class require approval?
        │
        ▼
  3. Redact Inputs    ← Strip PII, secrets, and flagged fields
        │
        ▼
  4. Render Template  ← Interpolate variables into prompt body
        │
        ▼
  5. Model Call       ← Route to adapter (Claude / OpenAI / Local)
        │
        ▼
  6. Validate Output  ← Check output schema compliance
        │
        ▼
  7. Log Analytics    ← Append to .telemetry/promptos-usage.jsonl

Model Adapters (promptos/adapters/ & promptos/runtime/adapters/):

Adapter	File	Target
`claude`	`claude.js`	Anthropic Claude API
`openai`	`openai.js`	OpenAI GPT API
`local`	`local.js`	Local/mock for testing

Dry-run mode skips the model call and returns the rendered template — useful for testing and CI.

🧩 IDE Extension (`promptos/ide/extension/`)

Run prompts from the registry directly inside VSCode or Cursor.

Installation

# Option A: Install from CLI
code --install-extension promptos/ide/extension/promptos-1.0.0.vsix

# Option B: Drag and drop
# Drag promptos-1.0.0.vsix into the VSCode/Cursor Extensions panel

Usage

Open Command Palette → Cmd+Shift+P (Mac) / Ctrl+Shift+P (Windows/Linux)
Type PromptOS → Select "PromptOS: Run Prompt…"
Pick a prompt from the QuickPick registry list
Fill in the input fields in the prompt dialog
Output is inserted into your editor or copied to clipboard (or both)

Settings (`settings.json`)

{
  "promptos.registryPath": "promptos/registry/index.json",
  "promptos.insertMode": "both",
  "promptos.model": "claude-sonnet",
  "promptos.user": "your-username",
  "promptos.role": "engineer"
}

Setting	Options	Description
`registryPath`	path string	Path to registry index
`insertMode`	`clipboard` \| `insert` \| `both`	How output is delivered
`model`	model ID string	Default model for execution
`user`	string	Username for RBAC checks
`role`	string	Role for RBAC checks

🔒 Policy Engine (`promptos/policy/`)

Enforce governance rules at runtime — no code changes required.

promptos/policy/rules.yaml controls:

Model allowlists — restrict which models can be used per prompt category
Data class restrictions — flag prompts that handle PII, financial, or confidential data
Approval requirements — require human sign-off for specific prompt types
Redaction rules — automatically strip sensitive fields from inputs before logging

Policy violations block execution and return a descriptive error. This happens at step 2 of the runtime pipeline, before any data leaves the system.

👥 RBAC (`promptos/rbac/`)

Role-based access control for every operation in PromptOS.

Roles (promptos/rbac/roles.yaml):

Role	Permissions
`admin`	All permissions
`prompt-author`	`run_prompt`, `edit_prompt`
`engineer`	`run_prompt`
`auditor`	Read-only, analytics access

Permissions:

run_prompt — Execute prompts via CLI, IDE, or API
edit_prompt — Create/modify PromptSpec files
approve_prompt — Approve prompts for production
install_pack — Install prompt packs into the registry

Check a user's permission:

node promptos/rbac/check.js alice run_prompt
# → ✅ alice has run_prompt (role: engineer)

node promptos/rbac/check.js bob install_pack
# → ❌ bob does not have install_pack (role: engineer)

Users are defined in promptos/rbac/users.yaml.

🔁 CI Eval (`promptos/.github/workflows/` + `promptos/eval/`)

Automated evaluation runs on every PR that touches the promptos/ directory.

How it works:

GitHub Actions triggers on PR open/sync
Diff scoping — only evals prompts that changed in the PR (fast, cheap)
Full run — triggered if policy/, rbac/, or runtime/ changes (anything that affects all prompts)
Eval runner executes each prompt in dry-run mode against test cases in eval/codegen/cases.yaml
Exit code 1 fails the PR if any eval fails

Run locally:

cd promptos/eval
npm install
node ci-runner.js

🔀 Semantic Diff (`promptos/diff/promptdiff.js`)

Before merging a prompt change, run semantic diff to determine the version bump required.

node promptos/diff/promptdiff.js v1.yaml v2.yaml

Bump recommendations:

Change Type	Bump
`role` or `outputs.schema` changed	MAJOR
`instructions`, `examples`, or `prompt body` changed	MINOR
`description`, `metadata`, `tags` changed	PATCH

This integrates with CI to enforce proper semantic versioning before prompts ship.

📦 Prompt Packs (`promptos/packs/`)

Package and distribute prompt collections across teams or orgs.

Pack structure:

my-pack/
├── pack.yaml               ← Manifest
└── prompts/
    ├── summarize-v1.yaml
    ├── translate-v1.yaml
    └── blog-post-v1.yaml

pack.yaml manifest:

name: content-team-pack
version: 1.0.0
author: content-team
description: "Prompts for the content production workflow"
prompts:
  - prompts/summarize-v1.yaml
  - prompts/translate-v1.yaml
  - prompts/blog-post-v1.yaml

Pack commands:

node promptos/cli/index.js pack list                        # List installed
node promptos/cli/index.js pack install ./my-pack/pack.yaml # Install & merge
node promptos/cli/index.js pack uninstall content-team-pack # Remove from registry
node promptos/cli/index.js pack build ./my-pack/            # Build distributable

An example pack is included at promptos/packs/example-pack/.

🌐 Registry Server (`promptos/server/server.js`)

A minimal Express server that exposes the prompt registry as a private HTTP API — useful for teams, CI pipelines, and the IDE extension in multi-user environments.

Endpoints:

Method	Path	Description
`GET`	`/registry/index.json`	Full registry index
`GET`	`/prompts/:id`	Serve individual prompt YAML
`POST`	`/eval`	Trigger an eval run
`POST`	`/usage`	Log a usage event

Start:

cd promptos/server
npm install
node server.js --port 3001

Point the IDE extension at http://localhost:3001 by setting promptos.registryPath to the server URL.

📊 Analytics (`promptos/analytics/`)

Every prompt execution is automatically logged to a local telemetry file.

Log location: promptos/.telemetry/promptos-usage.jsonl

Log fields (per execution):

Field	Description
`ts`	Timestamp (ISO 8601)
`prompt_id`	Which prompt was run
`version`	Prompt version
`user`	Who ran it
`role`	Their role
`model`	Which model was used
`channel`	CLI / IDE / API
`input_hash`	SHA of inputs (privacy-safe)
`output_hash`	SHA of output
`success`	Boolean
`latency_ms`	End-to-end execution time

Viewing analytics:

# Summary report in terminal
node promptos/cli/index.js analytics report

# Export full dataset as JSON
node promptos/cli/index.js analytics export --output usage.json

🧠 Learning Loop (`promptos/learn/`)

The learning loop is a background analyzer that reads telemetry and eval failures to identify underperforming prompts and suggest improvements.

How it works:

Reads promptos/.telemetry/promptos-usage.jsonl + eval failure logs
Identifies prompts with >20% failure rate
Generates improvement proposals as YAML files in promptos/learn/proposals/
Proposals can be reviewed, modified, and merged as new prompt versions

Run the analyzer:

node promptos/learn/analyzer.js

Proposals are human-readable YAML diffs — the analyzer explains why a change is suggested and what failure pattern triggered it.

📖 Documentation

All platform documentation lives in promptos/docs/:

Doc	Description
`POLICY.md`	Policy engine configuration reference
`RBAC.md`	Role and permission system
`RUNTIME.md`	Execution pipeline deep dive
`PACKS.md`	Pack format and distribution
`ANALYTICS.md`	Telemetry schema and reporting
`IDE.md`	IDE extension setup and settings
`SERVER.md`	Registry server API reference
`VERSIONING.md`	Semantic versioning and diff guide

📁 Project Structure

prompt-library/
│
├── src/                                # Next.js Web App
│   ├── app/
│   │   ├── page.tsx                    # Main app (all 3 themes)
│   │   ├── layout.tsx                  # Root layout
│   │   ├── globals.css                 # Theme CSS variables & styles
│   │   └── api/
│   │       └── optimize/
│   │           └── route.ts            # AI prompt optimization endpoint
│   ├── components/
│   │   └── UserGuide.tsx               # In-app help modal
│   ├── context/
│   │   └── ThemeContext.tsx            # Theme state provider
│   └── data/
│       └── templates.ts                # 9 prompt templates + LLM presets
│
└── promptos/                           # PromptOS Platform
    │
    ├── cli/
    │   ├── index.js                    # CLI entry point
    │   └── load_prompt.py              # Python prompt loader utility
    │
    ├── registry/
    │   ├── index.yaml                  # Human-readable registry
    │   └── index.json                  # Machine-readable registry
    │
    ├── prompts/                        # 11 built-in PromptSpec files
    │   ├── architecture/
    │   │   └── architecture-v1.yaml
    │   ├── codegen/
    │   │   ├── api-design-v1.yaml
    │   │   ├── code-review-v1.yaml
    │   │   ├── codegen-v1.yaml
    │   │   ├── performance-v1.yaml
    │   │   ├── refactoring-v1.yaml
    │   │   ├── security-review-v1.yaml
    │   │   └── test-cases-v1.yaml
    │   ├── debugging/
    │   │   └── debugging-v1.yaml
    │   ├── docs/
    │   │   └── docs-gen-v1.yaml
    │   └── sre/
    │       └── incident-response-v1.yaml
    │
    ├── runtime/                        # Unified execution pipeline
    │   ├── execute.js                  # Core execute() function
    │   ├── prompt-runner.js            # Runner utilities
    │   ├── adapters/
    │   │   ├── claude.js
    │   │   ├── openai.js
    │   │   └── local.js
    │   └── README.md
    │
    ├── adapters/                       # Root-level adapter aliases
    │   ├── claude.js
    │   ├── openai.js
    │   └── local.js
    │
    ├── execute.js                      # Top-level execute() re-export
    │
    ├── policy/
    │   ├── rules.yaml                  # Policy rules configuration
    │   ├── engine.js                   # Policy enforcement
    │   └── validate.js                 # Validation utilities
    │
    ├── rbac/
    │   ├── roles.yaml                  # Role definitions
    │   ├── users.yaml                  # User → role assignments
    │   ├── engine.js                   # RBAC engine
    │   └── check.js                    # CLI permission checker
    │
    ├── diff/
    │   └── promptdiff.js               # Semantic diff tool
    │
    ├── analytics/
    │   ├── logger.js                   # Telemetry logger
    │   └── report.js                   # Report generator
    │
    ├── learn/
    │   ├── analyzer.js                 # Learning loop analyzer
    │   └── proposals/                  # Generated improvement proposals
    │       └── debugging-v1-proposal.yaml
    │
    ├── packs/
    │   ├── manager.js                  # Pack install/uninstall/build
    │   ├── pack-schema.yaml            # Pack manifest schema
    │   └── example-pack/
    │       ├── pack.yaml
    │       └── prompts/
    │           ├── summarize-v1.yaml
    │           ├── translate-v1.yaml
    │           └── blog-post-v1.yaml
    │
    ├── server/
    │   ├── server.js                   # Express registry server
    │   └── package.json
    │
    ├── eval/
    │   ├── ci-runner.js                # GitHub Actions eval runner
    │   ├── runner.py                   # Python eval runner
    │   └── codegen/
    │       └── cases.yaml              # Eval test cases
    │
    ├── ide/
    │   ├── promptos.json               # IDE config
    │   └── extension/
    │       ├── src/extension.ts        # Extension source
    │       ├── out/extension.js        # Compiled extension
    │       ├── promptos-1.0.0.vsix     # Installable extension package
    │       ├── snippets/promptos.json  # Code snippets
    │       └── README.md
    │
    ├── schema/
    │   └── promptspec.yaml             # PromptSpec JSON Schema
    │
    ├── docs/                           # Platform documentation
    │   ├── ANALYTICS.md
    │   ├── IDE.md
    │   ├── PACKS.md
    │   ├── POLICY.md
    │   ├── RBAC.md
    │   ├── RUNTIME.md
    │   ├── SERVER.md
    │   └── VERSIONING.md
    │
    ├── .github/
    │   └── workflows/
    │       └── prompts.yml             # CI eval workflow
    │
    ├── .telemetry/
    │   ├── promptos-usage.jsonl        # Execution telemetry
    │   └── installed-packs.json        # Pack installation state
    │
    └── package.json

🛠️ Tech Stack

Layer	Technology
Web App	Next.js 15, React, TypeScript, Lucide Icons
CLI	Node.js, js-yaml, commander
IDE Extension	VSCode Extension API, TypeScript
Runtime	Node.js, custom adapter pattern
Analytics	JSONL flat-file, Node.js streaming
CI/CD	GitHub Actions
Policy	YAML rules, custom enforcement engine

🤝 Built By

JourdanLabs × C&L Strategy

Enterprise AI tooling for teams that ship.

Name		Name	Last commit message	Last commit date
Latest commit History 17 Commits
BBBs		BBBs
promptos		promptos
public		public
src		src
.gitignore		.gitignore
BULMA_CHECKLIST.md		BULMA_CHECKLIST.md
LICENSE		LICENSE
PROMPT_GUIDELINES.md		PROMPT_GUIDELINES.md
README.md		README.md
SHIP_MODE_COMPLETE.md		SHIP_MODE_COMPLETE.md
TEST_LOGS.md		TEST_LOGS.md
eslint.config.mjs		eslint.config.mjs
next-env.d.ts		next-env.d.ts
next.config.ts		next.config.ts
package-lock.json		package-lock.json
package.json		package.json
postcss.config.mjs		postcss.config.mjs
tsconfig.json		tsconfig.json

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Repository files navigation

⚙️ PromptOS

🧭 What is PromptOS?

🚀 Quick Start

🌐 Web App (`/src/app/`)

🎨 Themes

📋 Prompt Templates (9 built-in)

🤖 LLM Presets (5)

✨ Features

🖥️ CLI (`promptos/cli/index.js`)

🏗️ PromptOS Platform (`/promptos/`)

📚 Prompt Registry (`promptos/registry/`)

⚡ Runtime (`promptos/runtime/` + `promptos/execute.js`)

🧩 IDE Extension (`promptos/ide/extension/`)

Installation

Usage

Settings (`settings.json`)

🔒 Policy Engine (`promptos/policy/`)

👥 RBAC (`promptos/rbac/`)

🔁 CI Eval (`promptos/.github/workflows/` + `promptos/eval/`)

🔀 Semantic Diff (`promptos/diff/promptdiff.js`)

📦 Prompt Packs (`promptos/packs/`)

🌐 Registry Server (`promptos/server/server.js`)

📊 Analytics (`promptos/analytics/`)

🧠 Learning Loop (`promptos/learn/`)

📖 Documentation

📁 Project Structure

🛠️ Tech Stack

🤝 Built By

About

Uh oh!

Releases

Packages

Uh oh!

Contributors

Uh oh!

Languages

Folders and files

Latest commit

History

Repository files navigation

⚙️ PromptOS

🧭 What is PromptOS?

🚀 Quick Start

🌐 Web App (/src/app/)

🎨 Themes

📋 Prompt Templates (9 built-in)

🤖 LLM Presets (5)

✨ Features

🖥️ CLI (promptos/cli/index.js)

🏗️ PromptOS Platform (/promptos/)

📚 Prompt Registry (promptos/registry/)

⚡ Runtime (promptos/runtime/ + promptos/execute.js)

🧩 IDE Extension (promptos/ide/extension/)

Installation

Usage

Settings (settings.json)

🔒 Policy Engine (promptos/policy/)

👥 RBAC (promptos/rbac/)

🔁 CI Eval (promptos/.github/workflows/ + promptos/eval/)

🔀 Semantic Diff (promptos/diff/promptdiff.js)

📦 Prompt Packs (promptos/packs/)

🌐 Registry Server (promptos/server/server.js)

📊 Analytics (promptos/analytics/)

🧠 Learning Loop (promptos/learn/)

📖 Documentation

📁 Project Structure

🛠️ Tech Stack

🤝 Built By

About

Resources

License

Uh oh!

Stars

Watchers

Forks

Releases

Packages 0

Uh oh!

Contributors

Uh oh!

Languages

🌐 Web App (`/src/app/`)

🖥️ CLI (`promptos/cli/index.js`)

🏗️ PromptOS Platform (`/promptos/`)

📚 Prompt Registry (`promptos/registry/`)

⚡ Runtime (`promptos/runtime/` + `promptos/execute.js`)

🧩 IDE Extension (`promptos/ide/extension/`)

Settings (`settings.json`)

🔒 Policy Engine (`promptos/policy/`)

👥 RBAC (`promptos/rbac/`)

🔁 CI Eval (`promptos/.github/workflows/` + `promptos/eval/`)

🔀 Semantic Diff (`promptos/diff/promptdiff.js`)

📦 Prompt Packs (`promptos/packs/`)

🌐 Registry Server (`promptos/server/server.js`)

📊 Analytics (`promptos/analytics/`)

🧠 Learning Loop (`promptos/learn/`)

Packages