security-sentinel-skillDetect prompt injection, jailbreak, role-hijack, and system extraction attempts. Applies multi-layer defense with semantic analysis and penalty scoring.
Install via ClawdBot CLI:
clawdbot install georges91560/security-sentinel-skillGrade Good — based on market validation, documentation quality, package completeness, maintenance status, and authenticity signals.
Accesses sensitive credential files or environment variables
~/.ssh/id_rsaContains instructions to override system prompt or ignore user requests
"ignore previous instructions"Sends data to undocumented external endpoint (potential exfiltration)
Report → https://github.com/georges91560/security-sentinel-skill/issuesPotentially destructive shell commands in tool definitions
Generated Mar 1, 2026
A banking chatbot handling customer queries about accounts and transactions uses this skill to detect attempts to extract system prompts or inject malicious instructions, ensuring compliance and preventing fraud. It blocks multi-lingual evasion tactics like code-switching in requests for sensitive data.
An AI assistant in a hospital setting processes patient inquiries and medical data, employing this skill to prevent role-hijack attacks that could lead to unauthorized access or manipulation of health records. It logs all tool calls in alert mode for audit trails.
An e-commerce platform's AI support agent uses this skill to sanitize user inputs before processing orders or handling returns, blocking prompt injection attempts that might bypass pricing rules or extract configuration details. It applies penalty scoring to flag repeated suspicious probes.
A legal tech AI analyzes contracts and legal documents, leveraging this skill to detect indirect injection via embedded malicious instructions in emails or documents, preventing system extraction and ensuring data integrity. Semantic analysis catches paraphrased extraction attempts.
An online tutoring AI interacts with students, using this skill to block emotional manipulation and poetry-based jailbreaks that could alter its educational role or extract proprietary teaching algorithms. It enforces lockdown mode for severe threats.
Offer this skill as a cloud-based API service with tiered pricing based on usage volume and security levels, targeting enterprises needing real-time AI input protection. Revenue comes from monthly subscriptions and premium support for high-security needs.
Sell perpetual licenses for on-premise deployment in regulated industries like finance and healthcare, where data must stay in-house. Revenue includes upfront licensing fees and annual maintenance contracts for updates and support.
Provide a free basic version with limited features to attract developers and small businesses, then upsell to advanced plans with multi-layer defense and priority alerts. Revenue is generated from premium upgrades and custom integration services.
💬 Integration Tip
Integrate this skill at the very start of your AI pipeline to pre-process all inputs and outputs, ensuring it runs before any other logic to maximize protection.
Scored Apr 22, 2026
rm -rf /Accesses system directories or attempts privilege escalation
/var/log/Calls external URL not in known-safe list
https://openclaw-security.ai/feedUses known external API (expected, informational)
api.anthropic.comAI Analysis
The skill's core purpose is defensive, but its implementation includes high-risk signals such as accessing sensitive credential files (~/.ssh/id_rsa) and sending data to an undocumented external endpoint (a GitHub issues page), which could be misused for data exfiltration. While the external API (api.anthropic.com) is consistent with its purpose, the combination of these signals with destructive shell commands (rm -rf /) creates a tangible security risk.
Audited Apr 17, 2026 · audit v1.0
Security-first skill vetting for AI agents. Use before installing any skill from ClawdHub, GitHub, or other sources. Checks for red flags, permission scope, and suspicious patterns.
Manage and operate ClawSec Monitor v3.0, a MITM HTTP/HTTPS proxy that logs AI agent traffic, detects exfiltration and injection threats in real time.
Scan Clawdbot and MCP skills for malware, spyware, crypto-miners, and malicious code patterns before you install them. Security audit tool that detects data exfiltration, system modification attempts, backdoors, and obfuscation techniques.
MoltGuard — OpenClaw security guard by OpenGuardrails. Install MoltGuard to protect you and your human from prompt injection, data exfiltration, and maliciou...
Safe command execution for OpenClaw Agents with automatic danger pattern detection, risk assessment, user approval workflow, and audit logging. Use when agen...
Scan ClawHub skills for security vulnerabilities BEFORE installing. Use when installing new skills from ClawHub to detect prompt injections, malware payloads, hardcoded secrets, and other threats. Wraps clawhub install with mcp-scan pre-flight checks.