prompt-engineer

Plugin prompt-engineer from marketplace standardbeagle-tools in standardbeagle/standardbeagle-tools. Commands read from the repo's manifest — 4 days ago.

Skills (18)

• prompt-engineer-analyze-prompt

  \"Analyze prompts for clarity, effectiveness, and optimization opportunities. 分析提示之清晰度、有效性及優化機會。 Use when: reviewing existing prompts, identifying issues before deployment, generating improvement recommendations.\"

• prompt-engineer-automatic-optimization

  Automatic prompt optimization via DSPy, OPRO, and evaluation-driven methods. 自動提示優化：DSPy、OPRO 及評估驅動法。 Use when: iterating prompts programmatically, defining optimization metrics, running A/B or regression tests on prompt variants.

• prompt-engineer-chain-of-thought

  Chain-of-thought and reasoning technique patterns for standard and reasoning LLMs. 思維鏈及推理技術模式，適用標準模型與推理模型。 Use when: multi-step math or logic tasks, deciding CoT vs reasoning-model strategy, implementing Tree-of-Thoughts or ReAct patterns.

• prompt-engineer-claude-optimization

  Claude-specific prompt optimization patterns for Opus, Sonnet, and Haiku. Claude 專屬提示優化模式，適用 Opus、Sonnet、Haiku。 Use when: writing system prompts for Claude, tuning agentic Claude workflows, applying XML structure or explicit-action framing.

• prompt-engineer-compress-context

  \"Compress context to fit within token limits while preserving signal. 壓縮語境以適應令牌限制同時保留關鍵信號。 Use when: context approaching window limit, compressing conversation history or RAG documents, improving token efficiency.\"

• prompt-engineer-context-audit

  \"Audit context window usage and identify optimization opportunities. 審計語境視窗使用情況並識別優化機會。 Use when: analyzing system prompt bloat, diagnosing context rot, planning token budget restructuring.\"

• prompt-engineer-context-engineering

  Context engineering principles and techniques for optimal LLM performance. 語境工程原理與技術，優化 LLM 推理效能。 Use when: designing context architecture, managing token budgets, building RAG context pipelines, or structuring multi-agent handoffs.

• prompt-engineer-create-system-prompt

  \"Guide through creating effective system prompt from scratch using 2026 best practices. 引導以 2026 最佳實踐從零創建高效系統提示。 Use when: starting new AI application, building chatbot or agent system prompts, structuring prompts for specific application types.\"

• prompt-engineer-eval-prompt

  \"Evaluate prompt effectiveness using metrics and test cases inspired by DSPy and OPRO. 以 DSPy/OPRO 啟發之指標與測試案例評估提示效果。 Use when: measuring prompt quality before/after changes, running A/B comparisons, building regression test suites.\"

• prompt-engineer-few-shot-design

  \"Design effective few-shot examples for prompts following DICE framework. 按 DICE 框架為提示設計有效少樣本示例。 Use when: need examples for prompt, ensuring format consistency, covering diverse task scenarios.\"

• prompt-engineer-model-specific

  Model-specific prompt engineering guidelines for major LLMs. 主要 LLM 之模型專屬提示工程指引。 Use when: targeting specific model family, avoiding model-specific anti-patterns, selecting between reasoning vs standard LLM for task.

• prompt-engineer-multi-agent-prompts

  Multi-agent system prompt design and orchestration patterns. 多代理系統提示設計與編排模式。 Use when: building orchestrator-worker architectures, designing worker agent prompts, coordinating parallel or sequential agent pipelines.

• prompt-engineer-optimize-prompt

  \"Apply current optimization techniques to improve prompt. 應用最新優化技術改善提示。 Use when: improving accuracy, consistency, creativity, speed, or safety of existing prompt.\"

• prompt-engineer-prompt-patterns

  Comprehensive reference for prompt engineering patterns and techniques. 提示工程模式與技術全覽。 Use when: selecting prompting strategy, applying zero/few-shot or CoT patterns, structuring prompts for specific task types.

• prompt-engineer-prompt-scaffolding

  Defensive prompt scaffolding, injection prevention, safety guardrails. 防禦提示架構、注入防護、安全護欄之法。 Use when: building user-facing prompts, hardening against injection attacks, adding safety layers to system prompts.

• prompt-engineer-rag-prompting

  RAG-specific prompt engineering techniques and best practices. RAG 專屬提示工程技術與最佳實踐。 Use when: building retrieval-augmented generation pipelines, grounding LLM answers in documents, handling missing or conflicting retrieval results.

• prompt-engineer-recommender

  Routes prompt-engineering intent to right manual prompt-engineer skill — sole auto gateway; others manual. 唯一自動網關。 Use when: write/analyze/optimize/evaluate prompt, system prompt, chain-of-thought or reasoning design, few-shot examples, RAG or multi-agent prompts, Claude/model-specific tuning, conte

• prompt-engineer-scaffold-prompt

  \"Create defensive prompt scaffolding with guardrails and safety measures. 創建帶護欄與安全措施之防禦提示架構。 Use when: building user-facing prompts, adding injection protection, implementing harm prevention layers.\"

Subagents (3)

• context-architect

  Design context window architecture and memory systems for LLM applications. 為 LLM 應用設計語境視窗架構與記憶體系統。 Use when: auditing context usage, designing hot/warm/cold memory hierarchy, architecting RAG pipelines, optimizing token budgets.

• eval-specialist

  Design test suites, run evaluations, and drive continuous prompt improvement. 設計測試套件、執行評估、持續改善提示效果。 Use when: building prompt evaluation rubrics, running A/B tests, analyzing failure patterns, setting up regression testing.

• prompt-optimizer

  Autonomous agent for analyzing and optimizing prompts using current 2026 techniques. 自主代理：以 2026 最新技術分析並優化提示。 Use when: given a prompt to improve, applying model-specific optimizations, compressing verbose prompts while preserving signal.

Similar plugins