“guardrails and AI safety”

Secure AI coding agents (Claude Code, Cursor, Codex, Copilot) with permission boundaries, secret protection, code review gates, and safe sandbox…

Configure the shared guard against catastrophic shell commands in local AI agents.

NVIDIA's runtime safety framework for LLM applications.

Pre-commit hooks, CI gates, lint rules, database constraints and branch protection.

Rendered guardrail framework: keep a repo's hard rules in one-rule-per-file GRL sources behind a compact router, and render only the rules a task needs.

Use when an AI coding agent is writing or refactoring production code and its known failure modes need to be held in check — over-engineering, silent scope…

Seamlessly transitions conversation from AI to human staff when user requests a representative, complexity exceeds AI scope, or safety guardrails trigger.

Stop an AI agent damaging your repo: PreToolUse hooks, permission deny rules, protected paths, verification gates.

Use when reviewing AI-generated code before merge or execution: plan AST/security checks, isolated behavior tests and evidence, with language-specific…

Wrap agent loops in deterministic gates, scoped permissions and stop rules.

Security-first workflow for ssh-mcp-pro tasks involving remote execution, sensitive files, credentials, service impact, and production safety.

Cognitive load guardrails.

Set up Claude Code hooks to block dangerous Git commands (push, reset --hard, clean, branch -D, etc.) before they execute.

Set up Claude Code permission rules that make dangerous git commands (push, reset --hard, clean, branch -D, checkout ., restore .) ask the user before they run.

Install a PreToolUse hook that mechanically blocks destructive git commands (push, reset --hard, clean -f, branch -D, checkout .) before they execute.

This skill should be used when designing, planning, implementing, or reviewing any system that involves LLM agents, tool use, prompt construction, or agentic…

LLM and cloud cost awareness — model tiering, token budgets, right-sizing, and when a cheaper model suffices.

A constraint you write into a brief for a subagent or a future self binds you first; authoring a frame can make it invisible to you.

Installs guardrails in this project's .claude/ (and its git worktrees) that keep Claude Code from reading, writing or running commands on paths outside a…

Use this skill when you need evidence-bounded safety policy, abuse categories, refusal/redirect behavior, privacy, escalation, and Human risk decisions…

Use when page, component, form, table, dashboard, responsive layout, or UI review work needs product-quality implementation guardrails.

Configure and test agent guardrails for input/output validation, tool-use constraints, and safety boundaries

Meta's 7-8B specialized moderation model for LLM input/output filtering.

Design spec with 98 rules for building CLI tools that AI agents can safely use.