A specialized AI agent model for security analysis, threat detection, and safety evaluation tasks. Based on Qwen3-4B, optimized for agent workflows requiring security awareness and risk assessment.
This skill has safety concerns that you should review before use. Some patterns were detected that may pose a risk.Safety score: 60/100.
KillerSkills scans all public content for safety. Use caution before installing or executing flagged content.
A quantized GGUF model (AgentDoG-Qwen3-4B) designed for security-focused agent workflows, including threat analysis, safety evaluation, and risk assessment tasks.
This skill provides access to a specialized 4B parameter model trained for security and safety use cases. The model excels at:
The model is available in multiple quantization levels (IQ1_S through Q6_K) to balance performance and quality based on your hardware constraints.
When using this security agent skill:
1. **Model Selection**: Choose an appropriate quantization level based on your needs:
- Q4_K_M or Q4_K_S: Recommended for most use cases (fast, good quality)
- Q5_K_M or Q6_K: Higher quality for critical security analysis
- IQ3_S or IQ3_M: Lower resource usage, acceptable quality
- IQ2_* or IQ1_*: Only for severely constrained environments
2. **Security Analysis Workflow**:
- Clearly define the security scope (code review, threat assessment, etc.)
- Provide complete context about the system or code being analyzed
- Request specific threat categories if applicable
- Ask for severity ratings and remediation recommendations
3. **Safety Evaluation Workflow**:
- Describe the content or behavior to evaluate
- Specify safety criteria or policies to check against
- Request risk scores and detailed justifications
- Ask for mitigation strategies when issues are found
4. **Agent Monitoring Workflow**:
- Provide agent interaction logs or behavior patterns
- Define acceptable behavior boundaries
- Request anomaly detection and flagging
- Ask for compliance verification against safety guidelines
5. **Output Format**: Request structured outputs with:
- Severity/risk levels (e.g., Critical, High, Medium, Low)
- Specific findings with evidence
- Actionable remediation steps
- Confidence scores when applicable
**Security Code Review:**
```
Using AgentDoG Security Agent, analyze this authentication function for security vulnerabilities. Check for: SQL injection, XSS, authentication bypass, and insecure session handling. Provide severity ratings and remediation steps.
```
**Threat Assessment:**
```
Evaluate this API endpoint design for potential security threats. Consider: input validation, rate limiting, authentication, authorization, and data exposure risks. Rate each finding by severity.
```
**Content Safety Check:**
```
Review this user-generated content for safety concerns including: harmful instructions, personal information exposure, malicious intent, and policy violations. Provide a risk score and specific concerns.
```
**Agent Behavior Monitoring:**
```
Analyze these AI agent interaction logs for safety compliance. Flag any attempts to: bypass restrictions, generate harmful content, access unauthorized resources, or exhibit deceptive behavior.
```
Leave a review
No reviews yet. Be the first to review this skill!
# Download SKILL.md from killerskills.ai/api/skills/agentdog-security-and-safety-agent/raw