“tool-use”的搜索结果

Hunt LLM/AI feature bugs — prompt injection, indirect injection, exfiltration via tool-use/markdown, ASCII smuggling, agentic AI security (OWASP Agentic Apps 2026, ASI01-ASI10). Patterns: direct injection ('ignore previous instructions'), indirect injection via documents/web pages/email the model reads, ASCII smuggling (Unicode Tags block U+E0000-U+E007F, invisible to humans, decoded by the model), tool-use exfiltration (model has fetch/browse tool, attacker injects OOB URL, model exfils chat history/secrets), markdown-image zero-click exfil, system-prompt extraction, IDOR-via-AI (cross-tenant data). Targets: chatbots, RAG, summarizers, agentic copilots, MCP tools. Detection: any LLM-backed endpoint, doc upload triggering AI processing, autonomous agent with tools. Validate: OOB/Collaborator callback for exfil, verbatim-reproducible system-prompt leak (run twice), verifiable cross-tenant leak or RCE. Confabulation is NOT a finding. Use when hunting AI features, chatbots, RAG, agentic systems, MCP.

中文简介正在准备中。当前显示原文说明。

Design and operate multi-agent orchestration patterns (ReAct loops, evaluator-optimizer, orchestrator-workers, tool routing) for LLM systems. Use when building or debugging agent workflows, tool-use loops, or multi-step task delegation; triggers: agentic, multi-agent, orchestration, ReAct, evaluator-optimizer, tool-use, handoff.

中文简介正在准备中。当前显示原文说明。

Document the tools an AI agent may call, including input rules, caller permissions, execution limits, and retry conditions. Produce a tool register and verification plan for reviewing registrations or MCP routing designs.

中文简介正在准备中。当前显示英语简介。

• Defining which tools an agent may call
• Reviewing a proposed tool registration
• Documenting timeout and retry rules

Guides Python agent development with PydanticAI, covering tool calls and structured outputs checked against defined data types. Explains service injection, provider switching, streaming responses, and tests that run without calling a real language model.

中文简介正在准备中。当前显示英语简介。

• Building agents that return validated structured data
• Connecting customer support agents to order data
• Testing agent logic without real model calls

Design agents for multi-turn tool use — sequential tool calls, result accumulation, error recovery, and complex task decomposition over multiple turns.

中文简介正在准备中。当前显示原文说明。

AI Agent 架构模式目录。提供 ReAct、Multi-Agent Conversation、Multi-Agent Workflow、 RAG Agent、Tool-Use Agent、Autonomous Agent 六种核心模式的参考架构、组件拓扑、 数据流、设计决策和框架无关的最佳实践。供 agent-dev-workshop Phase 2 调用。

中文简介正在准备中。当前显示原文说明。

LLM and AI application security testing — prompt injection, jailbreak resistance, OWASP LLM Top 10 (2025), RAG and agent/tool-use security, model supply chain, and AI red teaming for authorized assessments

中文简介正在准备中。当前显示原文说明。

Evaluate Claude skill quality through auditing. Use when reviewing or auditing skills.

中文简介正在准备中。当前显示原文说明。

Autonomous AI agent tool router register: schema registration, runtime parameter coercion, idempotency keys, and error recovery policies.

中文简介正在准备中。当前显示原文说明。

Build production-ready AI agents with PydanticAI — type-safe tool use, structured outputs, dependency injection, and multi-model support.

中文简介正在准备中。当前显示原文说明。

Design reproducible evaluations for AI agents with representative task sets, explicit rubrics, appropriate graders, baselines, regression gates, and failure analysis. Use when defining agent quality, comparing prompts or models, validating a release, measuring tool-use reliability, investigating regressions, or deciding whether an agent is ready for production.

中文简介正在准备中。当前显示原文说明。

当需要为 AI 定义工具接口、设计调用规范、实现工具发现与编排机制时调用此 skill。典型场景包括：设计 AI agent 的工具集、定义 JSON Schema/XML/TypeScript 格式的工具描述、实现工具权限控制与并行调度、设计子代理委托架构。 不适用于：定义 AI 身份（persona-design）、安全规则（safety-guardrails）、记忆存储（memory-system）。 关键 trigger 信号：AI 需要调用外部 API 或服务、工具数量超过 10 个需要发现机制、需要区分工具权限层级、需要子代理并行执行、工具调用涉及敏感操作需要审批。

中文简介正在准备中。当前显示原文说明。

Integrate Claude AI into applications with the Anthropic SDK. Use when a user asks to add Claude to an app, use Claude for text generation, build a chatbot with Claude, use Claude's long context window, implement tool use with Claude, stream Claude responses, use Claude for code generation, document analysis, or reasoning tasks. Covers Messages API, streaming, tool use, vision, system prompts, extended thinking, and batch processing.

中文简介正在准备中。当前显示原文说明。

Expert in designing and building autonomous AI agents. Masters tool use, memory systems, planning strategies, and multi-agent orchestration. Use when "build agent, AI agent, autonomous agent, tool use, function calling, multi-agent, agent memory, agent planning, langchain agent, crewai, autogen, claude agent sdk, ai-agents, langchain, autogen, crewai, tool-use, function-calling, autonomous, llm, orchestration" mentioned.

中文简介正在准备中。当前显示原文说明。

Create high-quality, model-aware system prompts for any LLM (Claude, GPT, Gemini, open-source, etc.). Use this skill whenever the user wants to create, write, build, design, draft, or improve a system prompt, system instructions, or custom instructions for any AI model. Also trigger when the user asks about prompt engineering, prompt design, prompt optimization, or wants to define behavior for an AI assistant, chatbot, agent, or any LLM-powered application — even if they don't explicitly say "system prompt". Covers all use cases including chatbots, agentic systems, tool-use workflows, content generation, data extraction, code assistants, and multi-turn conversations.

中文简介正在准备中。当前显示原文说明。

Conversational guide for using tools/session-retro/observe.mjs to capture friction, wins, corrections, decisions, gaps, tool-use, and checkpoints as they happen during a session. Use at session start (to kick off the log), at inflection points (to record events), and at session end (to archive and feed the retrospective generator).

中文简介正在准备中。当前显示原文说明。

Create, improve, or audit Vapi voice agent and Squad system prompts for production phone and web based voice agents. Use when the user wants help designing a Vapi assistant prompt, multi-assistant Squad prompt set, refining an existing prompt, creating prompt sections, building an intake or handoff workflow, improving tool-use instructions, adding guardrails, or optimizing voice-agent behavior for brevity, turn-taking, error handling, caller data collection, escalation, handoffs, and spoken formatting.

中文简介正在准备中。当前显示原文说明。

Design tool-use patterns for AI agents. TRIGGERS - Use when user needs help with ai-agent-tools related tasks.

中文简介正在准备中。当前显示原文说明。

AI and LLM red-teaming skill for evaluating prompt injection, jailbreak, data exfiltration, tool abuse, agent hijack, RAG poisoning, model denial of service, and unsafe tool-use chains in Gemini, Claude, GPT, and open-weight models. Use to build evaluation harnesses, attack corpora, defensive guardrails, and red-team reports for AI systems you own or are authorized to test.

中文简介正在准备中。当前显示原文说明。

Provider-agnostic patterns for LLM function calling, tool loops, and agentic workflows

中文简介正在准备中。当前显示原文说明。

Deep API security assessment beyond surface scanning. Covers the full OWASP API Security Top 10 (2023): Broken Object Level Authorization (BOLA / IDOR), Broken Authentication, Broken Object Property Level Authorization (mass assignment + excessive data exposure), Unrestricted Resource Consumption, Broken Function Level Authorization (BFLA / vertical privilege escalation), Unrestricted Access to Sensitive Business Flows, Server-Side Request Forgery via API parameters, Security Misconfiguration, Improper Inventory Management (shadow/zombie/deprecated endpoints, v1/v2 drift), and Unsafe Consumption of third-party APIs. Works across REST, GraphQL, gRPC, SOAP, and MCP servers. Discovers APIs from OpenAPI/Swagger specs, GraphQL introspection, gRPC reflection, .well-known endpoints, JS bundles, and traffic capture. Uses kiterunner, ffuf, schemathesis, restler-fuzzer, openapi-fuzzer, graphql-cop, clairvoyance, batchql, inql, jwt_tool, postman, mitmproxy, and manual http(action="request", ...) payloads. Every technique includes actual payloads, commands, and verification logic. Chains from /pentester or /codebase when API endpoints are discovered, chains into /web-exploit when classic injection points are found in API parameters, chains into /post-exploit when RCE is achieved, and chains into /ai-redteam when an LLM/AI endpoint is discovered (chat APIs, completion endpoints, RAG search, agentic tool-use, MCP servers).

中文简介正在准备中。当前显示原文说明。

Fundamental patterns for effective agentic behavior. Teaches decomposition, tool orchestration, error recovery, context management, quality self-assessment, and knowing when to stop. Model-agnostic principles that make any agent more effective regardless of domain. Activate on: "how should I structure this agent", "agentic workflow", "agent patterns", "multi-step task", "tool orchestration", "/agentic-patterns", "decompose this", "agent best practices", "chain of actions", "when should the agent stop", "agent loop design". NOT for: creating agent infrastructure (use agent-creator), building DAGs (use windags-architect), specific tool implementation.

中文简介正在准备中。当前显示原文说明。

Scientific model and eval tuning loop for prompt/skill/system-prompt changes, model configuration, runtime options, and local/Ollama model sizing or environment variables. Use when tuning eval performance, comparing local or remote models, improving tool-use behavior, reducing timeouts/loops, or deciding whether an eval optimization is worth keeping.

中文简介正在准备中。当前显示原文说明。

Construa agentes de IA prontos para produção com PydanticAI — type-safe tool use, structured outputs, dependency injection, e suporte multi-modelo.

中文简介正在准备中。当前显示原文说明。