mirror of
https://github.com/BagelHole/DevOps-Security-Agent-Skills.git
synced 2026-08-22 12:49:53 +02:00
39 lines
1.1 KiB
Markdown
39 lines
1.1 KiB
Markdown
|
---
|
||
|
|
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
|
||
|
|
|
||
|
|
1. Isolate tool execution with strict allowlists.
|
||
|
|
2. Add policy checks before sensitive actions.
|
||
|
|
3. Limit token scope and credential lifetimes.
|
||
|
|
4. Apply output filtering for sensitive data.
|
||
|
|
5. 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](../llm-app-security/) - Application-layer LLM defenses
|
||
|
|
- [threat-modeling](../operations/threat-modeling/) - Structured risk analysis
|