By name: ai-agent-security description: Secure AI agents against prompt injection, tool abuse, and data exfiltration with defense-in-depth controls. license: MIT metadata: author: devops-skills version: "1.0"
AI Agent Security
Protect agentic systems from adversarial input and unsafe tool execution.
Threats to Model
- Prompt injection through untrusted content
- Excessive permissions on tools and APIs
- Data exfiltration via model responses
- Cross-tenant context leakage
Security Controls
- Isolate tool execution with strict allowlists.
- Add policy checks before sensitive actions.
- Limit token scope and credential lifetimes.
- Apply output filtering for sensitive data.
- Log every privileged tool invocation.
Incident Readiness
- Keep immutable audit trails for prompts and tool calls.
- Build kill switches for high-risk tools.
- Run regular red-team scenarios.
Related Skills
- llm-app-security - Application-layer LLM defenses
- threat-modeling - Structured risk analysis