38 skills · sorted by relevance
Configure safety checks for LLM applications with NeMo Guardrails. Covers attempts to bypass safety rules, input and output validation, masking personal information, and checking factual claims through programmable rules and integrations.
Add safety checks to LLM applications with NeMo Guardrails. This skill guides input and output validation, detection of attempts to bypass safety rules, personal information masking, and checks for unsupported or inaccurate answers.
Add checks to LLM prompts and responses using Llama Guard 3, NeMo Guardrails, and LLM Guard. Configure content screening, conversation boundaries, and sensitive-data filtering, then measure blocking and false-positive rates with labeled test prompts.
Reviews Rust applications and libraries for security issues involving unsafe code, memory safety, concurrency, calls across language boundaries, panics, and async runtimes. Coordinates specialized reviewers and produces Markdown and SARIF reports, a format for analysis tools.
Use for AI-assisted Unity work that needs live repo discovery, project-derived routing, runtime-owner proof, runtime-visible output hard stops, runtime numeric proof for repeated visible-output failures, state-step guards, multi-agent scope ownership, modular C#/asmdef safety, UI/scene/visual asset gates, data-first content changes, validation, cleanup proof, or durable workflow rules. Best when agents must prove the actual folder/module/scene/prefab/runtime owner before editing, especially runtime UI, generated assets, code graphs, tutorial/state flows, guided selection/action flows, overlay/dim source-bound mistakes, coordinate conversions, focus/highlight/marker/HUD alignment, predictive lead aiming, ballistic trajectories, spatial IoU verification, or repeated "fix still not visible" failures.
The English summary is being prepared. Showing the original description.
Create Terraform files that define alerts for AI agent latency, errors, token use, safety, security, and response quality. Check telemetry and existing policies first, then validate the generated configuration. Quality alerts require Vertex AI.
Review existing DeerFlow skill packages for publication readiness, activation rules, safety boundaries, supporting files, and evaluation evidence. Receive concrete findings and suggested replacement text, with a clear distinction between inspection results and verified runtime behavior.
Plan and implement infrastructure, containers, CI/CD, deployment safety, observability, and rollback changes. Use for Docker, IaC, cloud runtime, Kubernetes, pipeline, deployment, or operational reliability work. Does not read dotenv files or approve/release deployments.
The English summary is being prepared. Showing the original description.
Framework de segurança em runtime da NVIDIA para aplicações LLM. Com detecção de jailbreak, validação de entrada/saída, verificação de fatos, detecção de alucinações, filtragem de PII, detecção de toxicidade. Usa DSL Colang 2.0 para rails programáveis. Pronto para produção, executa em GPU T4.
The English summary is being prepared. Showing the original description.
Operational safety for deployless repositories. Use after mainline deployment and runtime release controls are designed, when a repo needs traces, replay tests, observability, rollback drills, smoke checks, compatibility windows, or expand-contract migration safety.
The English summary is being prepared. Showing the original description.
Review TypeScript settings and indexed access to arrays and objects to find values that may be undefined. Enable compiler checks and update unsafe code so potential failures become visible before the code runs.
Design TypeScript types and APIs that catch mistakes such as mixed-up identifiers and missing state handling. Implement reusable types and runtime checks that narrow types, configure tRPC for shared client-server types, and improve compiler settings.
Verify code correctness statically against specifications using type checking, contract verification, and formal methods. Use when: (1) Verifying type safety and null safety in Python, Java, or C/C++ code, (2) Checking design-by-contract specifications (preconditions, postconditions, invariants), (3) Validating code against formal specifications, (4) Ensuring code quality and correctness before runtime, (5) Finding potential bugs through static analysis. Supports Python (mypy, contracts), Java (javac, JML), and provides verification scripts and contract specification guidelines.
The English summary is being prepared. Showing the original description.
Performs abstract interpretation over source code to infer possible program states, variable ranges, and data properties without executing the program. Reports potential runtime errors including out-of-bounds accesses, null dereferences, type inconsistencies, division by zero, and integer overflows. Use when analyzing code for potential runtime errors, performing static analysis, checking safety properties, or verifying program behavior without execution.
The English summary is being prepared. Showing the original description.
Expert guide for AI Safety, Governance, and Responsible AI in production — Constitutional AI enforcement, runtime guardrails (NeMo Guardrails 2.0, Llama Guard 3), bias auditing, hallucination detection, EU AI Act compliance, model cards, and content provenance / Panduan ahli untuk Keamanan AI, Tata Kelola, dan AI Bertanggung Jawab di produksi.
The English summary is being prepared. Showing the original description.
Provides TypeScript-specific code review patterns covering strict mode, ESM, type safety, branded types, discriminated unions, async patterns, runtime safety, and common anti-patterns. Activates on detection of tsconfig.json, *.ts, or *.tsx files during code review — loaded automatically by dh:code-reviewer.
The English summary is being prepared. Showing the original description.
Configures best-practice alerting policies for AI agents using OpenTelemetry (OTel) metrics, generating output as Terraform (.tf) configuration files. Use when analyzing, writing, or deploying alerting policies to monitor agent latency, error rates, token usage, and quality metrics. Don't use for standard infrastructure monitoring unrelated to AI agents, or when the agent is not instrumented with OpenTelemetry (for Reliability, Cost, Safety, Security alerts). NOTE: Reliability, Cost, Safety, and Security alerts use generic OTel metrics and work across runtimes (such as Cloud Run, Vertex AI). Quality alerts rely on Vertex AI Online Monitors and are strictly bound to Vertex AI deployments.
The English summary is being prepared. Showing the original description.
Designs human-in-the-loop review points for DAG workflows. Determines what to present to the human, how to collect feedback, and how to route approve/reject/modify decisions back into the DAG. Use when adding approval gates, designing review UX, or handling human feedback in agent workflows. Activate on "human review", "approval gate", "human-in-the-loop", "human gate", "approval workflow", "user review step". NOT for executing human gates at runtime (use dag-runtime with Temporal signals), general UX design, or chatbot conversation design.
The English summary is being prepared. Showing the original description.
Manage continuously running or cloud-hosted agents with lifecycle controls, logs, metrics, restricted permissions, and emergency stops. It guides incident response, staged deployments, rollbacks, and tracking success rates and cost per successful task.
Review JavaScript JSON parsing for malformed input and unexpected data structures. Add error handling and validate parsed data before use, helping prevent failures when reading API responses, browser storage, or user-provided content.
Plan and review fluid simulations with FluidSim, from solver selection and resource estimates to output analysis and restart checks. Examine numerical accuracy through grid and time-step comparisons, conservation checks, and explicit limits on computing resources.
Turn a validated build sheet into four API request files and a resumable launch script for Claude Managed Agent. Check the files for exposed keys before users launch with their own Anthropic key.
Provides guidance for managing Twitter/X content, including weekly plans, posts, threads, replies, and performance analysis. Includes API v2 examples and review queue templates to help organize publishing and maintain a consistent brand voice.
Add, review, or fix JavaScript interop in Blazor components. USE FOR: calling JavaScript from Blazor, calling .NET from JavaScript, collocated .razor.js modules, IJSRuntime, IJSObjectReference lifecycle, DotNetObjectReference, ElementReference, timing rules for when JS is available, IAsyncDisposable disposal of JS references, server-side JS interop safety. DO NOT USE FOR: general Blazor component authoring without JS interop needs (use author-component), forms (use collect-user-input).
The English summary is being prepared. Showing the original description.