Maintainer of this skill?
Automated security scan — not a certification.
by Rkcr7
An undetectable AI assistant that lives on your screen. Solves problems, answers questions, analyzes screenshots — invisible to everyone but you.
# Add to your Claude Code skills
git clone https://github.com/Rkcr7/Aura-AIGuides for using ai agents skills like Aura-AI.
Last scanned: 11/7/2026
{
"issues": [],
"status": "PASSED",
"scannedAt": "2026-11-07T10:54:27.619Z",
"npmAuditRan": true,
"pipAuditRan": false,
"promptInjectionRan": true
}See how Aura-AI compares with popular alternatives.
Aura-AI is an open-source ai agents skill for AI coding assistants such as Claude Code, Codex CLI, and ChatGPT, built by Rkcr7. An undetectable AI assistant that lives on your screen. Solves problems, answers questions, analyzes screenshots — invisible to everyone but you. It has 100 GitHub stars.
Yes. Aura-AI passed SkillsLLM's automated security scan — a dependency vulnerability audit plus prompt-injection heuristics — with no high-severity issues. You can read the full report in the Security Report section on this page.
Clone the repository with "git clone https://github.com/Rkcr7/Aura-AI" and add it to your Claude Code skills directory (see the Installation section above).
Aura-AI is primarily written in JavaScript. It is open-source under Rkcr7 on GitHub, so you can review or fork the full source.
Yes. SkillsLLM lists many other AI Agents skills you can browse and compare side by side. Open the AI Agents category from the badge at the top of this page, or use the Related Skills and comparison links further down to weigh Aura-AI against similar tools.
No comments yet. Be the first to share your thoughts!
From $49 · 30 days · terms & refunds
⚠️ Third-Party Software Notice
This skill is third-party open-source software developed and hosted independently on GitHub. SkillsLLM is an informational directory and does not control or maintain the underlying repository.
Any security checks, ratings, or warnings displayed by SkillsLLM are automated and limited in scope. They do not constitute a security certification or guarantee that the software is safe, error-free, or free from malicious code, vulnerabilities, compromised dependencies, or prompt-injection risks.
Review the source code, permissions, dependencies, and configuration before installing or running any third-party skill. Use is at your own risk. To the maximum extent permitted by applicable law, SkillsLLM is not liable for losses arising from third-party software.
Automated security scan — not a certification.
An invisible AI overlay that lives on your screen — answers questions, solves problems, analyzes screenshots, and feeds you real-time intelligence. Works during interviews, exams, meetings, or anything on your screen. Invisible to screen recordings. Undetectable by proctoring software. You see everything. They see nothing.
👻 Invisible by Design — Screen capture protected. Hidden from taskbar. Undetectable by Zoom, Teams, Google Meet, and every proctoring tool on the market.
💸 $0. Forever. — No subscriptions. No credits. No paywalls. Bring your own free API keys and own it.
⚡ Stupidly Fast — Powered by Cerebras (~2,000-3,000 tokens/sec) and Groq (~400-700 tokens/sec) — the two fastest AI inference engines on Earth. Answers hit your screen before the interviewer finishes talking. Nothing else comes close.
🖥️ Platform Support — Windows 10/11 is the primary, fully-supported platform (this
masterbranch). A macOS port is in active development on themacosbranch — feature parity is work-in-progress and some stealth behaviors are limited compared to Windows. macOS contributors welcome via PRs targeting themacosbranch.
| ⚡ Opens in about 2 seconds | The window shows up right away with an Aura AI splash, and the app is ready in ~1.9s instead of 5 to 9s |
| 📴 No internet needed to load | Fonts, syntax highlighting and math ship with Aura now, so the UI loads the same offline or on a network that blocks Google Fonts or cdnjs |
| 🧮 Math renders properly | LaTeX in answers shows up as real equations instead of raw $...$ |
| 🩺 Blank window explained | If WebView2 can't start, the terminal says why and how to fix it instead of leaving an empty window |
| 🔌 Starts without internet | run.bat only runs pip when requirements.txt changes, so it launches in ~2s and no longer stalls for minutes offline |
| ⚡ Cerebras Vision | Cerebras now does image input via Gemma 4 31B at ~1850 tok/s — the fastest screenshot-to-answer path Aura has ever had, and the new vision default |
| 🚀 Latest models | Refreshed across every provider: GPT-OSS 120B, Gemma 4 31B, Gemini 3.6 / 3.5 Flash |
| 🎯 Tuned for speed | Cerebras drives both text and vision by default. Need maximum accuracy on a dense screenshot? Alt+T swaps to Gemini 3.6 Flash |
| 🏷️ Labelled model picker | Every model now shows its speed/accuracy tradeoff right in the dropdown, so you pick the right one under pressure |
| 🎤 Leaner audio pipeline | ~70% less bandwidth and 8× fewer messages, with sample rate locked to Deepgram for cleaner transcription |
| 👁️ Sharper at low opacity | Headings, bold text and code blocks are noticeably easier to read at 40% transparency |
| ♿ Respects reduced motion | Animations settle when your OS asks for it — and a stiller overlay is a less noticeable one |
Upgrading? Re-copy
ai_providers.example.jsonto pick up the new model lineup — provider catalogues moved on since v1.0.0.
Aura isn't just another interview prep tool. It's your secret weapon for any challenge—from aptitude tests to quantitative brain-twisters, behavioral showdowns to certification exams. Aura is a revolutionary AI assistant that operates in real-time, providing candidates with the critical insights they need to excel in high-stakes situations—all without ever tripping tab-switching warnings, thanks to its stealthy, seamless design.
Imagine having a world-class expert whispering in your ear, helping you deconstruct complex problems, articulate your thoughts, and navigate the toughest questions with confidence. That's Aura.
Aura sits on top of your screen as an invisible overlay during live interviews and exams. It listens to conversations in real-time via Deepgram speech-to-text, generates intelligent coaching responses using blazing-fast LLMs (Groq, Cerebras, Gemini), and lets you screenshot any problem for instant Vision AI solutions—all while staying completely hidden from screen recordings, proctoring software, and human observers.
Real screenshots and recordings from a live Aura session — no mockups, no staging.
Fill in your name, target company, and role, then select your interview focus (Behavioral, Coding, System Design, Data Structures, Cultural Fit, or HR/Screening). Paste your full resume and job description for hyper-personalized coaching. Hit the ⚡ Demo button to auto-fill with sample data for quick testing.
Choose your Primary AI Model (default for all responses — Cerebras shown here for maximum speed) and an optional Secondary Model (switch anytime with Alt+W). Below, configure your Vision AI Model for screenshot analysis — Gemini delivers the best accuracy, Groq Llama 4 is the fastest. The Quick Switch Hotkeys panel shows Alt+Q, Alt+W, Alt+E for instant model switching mid-interview.
The Advanced Config tab is the raw control panel. Enter your Deepgram API key (speech-to-text), then manage multiple API keys per provider — notice Groq showing 5 keys for maximum rate-limit resilience. Each provider shows its base URL, keys, and a Test Connection button to verify everything works before going live. Changes are saved directly to ai_providers.json and .env.
The final onboarding tab is your cheat sheet. Three highlighted pro-tip cards cover: 🔒 Stealth Professional Tip (always use global hotkeys — never click Aura during proctored exams), 📸 Multi-Screenshot Analysis (queue up to 4 screenshots with Alt+S, process all at once with Alt+P), and ⚡ Speed Strategy (use Alt+E to auto-select the fastest healthy provider). Below, every hotkey is organized by category with color-coded badges.
Microphone Permission Request — Aura's Pre-Flight Check runs before every session. It validates your microphone access, backend connection, Deepgram API, and all configured AI providers. Grant the mic permission prompt to proceed.
All Systems Go ✅ — Every check is green: Microphone Permission OK, Microphone Selected (NVIDIA Broadcast detected), Backend Connected, Deepgram API OK, Primary Cerebras OK, Secondary Groq OK, Primary Vision Gemini OK, Secondary Vision Groq OK. Hit Start Interview and you're live.
See how Aura overlays directly on top of LeetCode during a live coding session — invisible to screen recordings, fully transparent, and controlled entirely via hotkeys.
Screen Share Protection — Aura runs during a live LeetCode session with "Share Entire Screen" active. The Aura window overlays the coding environment but is excluded from the screen capture via WDA_EXCLUDEFROMCAPTURE. Notice the "Microphone Muted" status bar and model switcher visible only to the user.
Maximum Transparency Mode — With Alt+1 (40% opacity), Aura becomes a near-invisible ghost overlay. The Live Interview Session indicator, MUTED/RESET controls, and AI model badges (Primary: Cerebras) are barely visible — just enough for you, completely invisible to anyone observing your screen. The LeetCode problem
