Skill Auditor Desktop — Security skill for Claude Code
Security audit skill for claude.ai / Claude Desktop — scans uploaded skills for prompt injection, data exfiltration, and privilege escalation before you trust them.
How to install Skill Auditor Desktop
This entry records only its repository, not the path inside it, so there is no
exact command to give. Open ki-stuff/skill-auditor-desktop and copy the folder into
~/.claude/skills/, or the file into ~/.claude/agents/.
What Skill Auditor Desktop does
Security audit skill for claude.ai / Claude Desktop — scans uploaded skills for prompt injection, data exfiltration, and privilege escalation before you trust them. Desktop edition of ki-stuff/skill-auditor.
Alternatives in Security
- Security Threat Model — Generate repo-specific threat models identifying trust boundaries 14.6k ★
- Synk Skill Security Scanner — Agent Trust Hub 1.9k ★
- Promptmap — Security scanner for custom LLM apps 1.2k ★
README
🛡️ Skill-Auditor Desktop
🇩🇪 Kurzbeschreibung
**Skills sind Prompts. Manche bringen Scripts mit. Beides kann feindlich sein — und beim Hochladen prüft das niemand.**
Der Skill-Auditor Desktop schließt genau diese Lücke für **claude.ai und Claude Desktop**: Lade den Skill, den du prüfen willst (Datei, Ordner oder `.zip`), in den Chat hoch und frag, ob er sicher ist. Der Auditor führt automatisch einen Muster-Scan plus eine inhaltliche Prüfung durch — gegen Prompt Injection, Datenexfiltration, Rechteausweitung, Persistenz und Social Engineering — und gibt dir eine einfache Ampel-Einschätzung (🟢 / 🟡 / 🔴) in Klartext zurück. Kein Kommando, kein Pfad, kein Fachjargon nötig.
**Nutzt du Claude Code (die CLI)?** Diese Fassung ist bewusst auf die claude.ai/Desktop-Sandbox zugeschnitten (kein Dateisystemzugriff, daher Upload statt Pfadangabe, kein Inventar-Modus). Für die CLI mit vollem Funktionsumfang — inklusive Dashboard über alle installierten Skills — die Schwesterfassung nutzen: [skill-auditor](https://github.com/ki-stuff/skill-auditor).
A Claude Skill that vets *other* skills before you trust them — built specifically for **claude.ai and Claude Desktop**, where Skills run in an isolated sandbox with no access to your local filesystem.
Upload the skill you want to check (a file, a folder, or a `.zip`) into the conversation and ask whether it's safe. The auditor runs an automated pattern scan plus a manual content review — prompt injection, data exfiltration, privilege escalation, persistence, social engineering — and replies with a plain-language traffic-light verdict (🟢 / 🟡 / 🔴). No commands, no paths, no jargon required.
**Using Claude Code (the CLI) instead?** Use the full-featured sibling project instead: [skill-auditor](https://github.com/ki-stuff/skill-auditor). It has direct filesystem access, so it can also scan everything you already have installed under `~/.claude/skills/` (inventory mode) — a claude.ai/Desktop Skil
Related Skills
Skillguard
Security scanner for AI agent skills. Detects prompt injection, data exfiltration, and malicious payloads befo
Skill Lint
Security scanner for Claude Code / agent skills. Catches prompt injection, obfuscation, credential exfiltratio
Skill Sentinel
A security scanner for Agent Skill packages. Skill Sentinel uses multi-agent AI analysis to detect prompt inje
Skillspector
Security scanner for AI agent skills. Detect vulnerabilities, malicious patterns, security risks, prompt injec
Auth Audit
Full auth & session audit: login flow, session/cookie security, JWT handling, middleware safety, privilege esc
Skill Auditor Seguranca
Executable security and LGPD gate for AI-built projects. Scans for exposed secrets, .env leaks, Supabase servi
Related Agents
Lens AI Security
AI security lens of the production readiness audit. Treats LLM and agent integrations as attack surface - prom
Sec Arch
Security-focused L2+ specialist for authn/authz, trust boundaries, privilege escalation, and dangerous API/sec
Ad Security Reviewer
Use this agent when you need to audit Active Directory security posture, evaluate privilege escalation risks,
