Apply multi-layer guardrails to AI agent pipelines against prompt injection and unsafe outputs.
npx skills add https://github.com/Sheldon-92/TAD --skill ai-guardrails
Please help me install this Agent Skill. Skill: ai-guardrails Source: https://github.com/Sheldon-92/TAD/tree/main/.agents/skills/ai-guardrails Command: npx skills add https://github.com/Sheldon-92/TAD --skill ai-guardrails
ai-guardrails addresses the need to harden AI agent pipelines against prompt injection, unsafe outputs, and uncontrolled agent behavior by providing a configurable, multi-layer guardrail framework.
Load the ai-guardrails pack into your agent workflow and run the included guardrail validator to identify and remediate risk findings.
High-intent search queries and answers about installing and using this skill.
Implement PII de-identification and output validation in LLM pipelines by applying deterministic guardrail validation scripts. This framework redacts sensitive data and verifies generated content before it reaches end-users in end-to-end LLM apps.
Enforce tool-call gating to restrict excessive agency in AI agents by configuring cross-cutting guardrails. This framework gates tool execution within the processing stage, preventing uncontrolled agent behavior in enterprise environments.
Audit AI guardrails configurations for OWASP 2025 risk mappings by running the included deterministic validator script. This enables repeatable security audits and supports risk discovery and remediation directly within your CI pipeline.
This defense-in-depth approach works with retrieval-augmented generation workflows by applying dedicated RAG safeguards. It ensures content moderation and deterministic output validation are enforced across input, processing, and output stages.
Deterministic scripts for AI output validation provide repeatable security audits but require manual configuration of guardrail rules. You must continuously update these deterministic checks to address evolving prompt injection techniques and unsafe outputs.
HyperscapeAIConfigure API gateways with rate limiting and JWT authentication.
jd-opensourceIdentifies and exploits AI/LLM security vulnerabilities in enterprise deployments using Garak and Burp Suite workflows.
passionseedConfigure and enforce content safety guardrails for TrueFoundry AI Gateway.
latestaiagentsValidate and sanitize user inputs to defend against prompt injection attacks.
