{ "schemaVersion": "1.0", "item": { "slug": "prompt-engineering-2", "name": "Prompt Engineering", "source": "tencent", "type": "skill", "category": "AI 智能", "sourceUrl": "https://clawhub.ai/kongyo2/prompt-engineering-2", "canonicalUrl": "https://clawhub.ai/kongyo2/prompt-engineering-2", "targetPlatform": "OpenClaw" }, "install": { "downloadMode": "redirect", "downloadUrl": "/downloads/prompt-engineering-2", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=prompt-engineering-2", "sourcePlatform": "tencent", "targetPlatform": "OpenClaw", "installMethod": "Manual import", "extraction": "Extract archive", "prerequisites": [ "OpenClaw" ], "packageFormat": "ZIP package", "includedAssets": [ "SKILL.md", "references/antipatterns.md", "references/evaluation.md", "references/examples.md", "references/patterns.md", "references/techniques.md" ], "primaryDoc": "SKILL.md", "quickSetup": [ "Download the package from Yavira.", "Extract the archive and review SKILL.md first.", "Import or place the package into your OpenClaw setup." ], "agentAssist": { "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.", "steps": [ "Download the package from Yavira.", "Extract it into a folder your agent can access.", "Paste one of the prompts below and point your agent at the extracted folder." ], "prompts": [ { "label": "New install", "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete." }, { "label": "Upgrade existing", "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run." } ] }, "sourceHealth": { "source": "tencent", "status": "healthy", "reason": "direct_download_ok", "recommendedAction": "download", "checkedAt": "2026-04-30T16:55:25.780Z", "expiresAt": "2026-05-07T16:55:25.780Z", "httpStatus": 200, "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network", "contentType": "application/zip", "probeMethod": "head", "details": { "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network", "contentDisposition": "attachment; filename=\"network-1.0.0.zip\"", "redirectLocation": null, "bodySnippet": null }, "scope": "source", "summary": "Source download looks usable.", "detail": "Yavira can redirect you to the upstream package for this source.", "primaryActionLabel": "Download for OpenClaw", "primaryActionHref": "/downloads/prompt-engineering-2" }, "validation": { "installChecklist": [ "Use the Yavira download entry.", "Review SKILL.md after the package is downloaded.", "Confirm the extracted package contains the expected setup assets." ], "postInstallChecks": [ "Confirm the extracted package includes the expected docs or setup files.", "Validate the skill or prompts are available in your target agent workspace.", "Capture any manual follow-up steps the agent could not complete." ] }, "downloadPageUrl": "https://openagent3.xyz/downloads/prompt-engineering-2", "agentPageUrl": "https://openagent3.xyz/skills/prompt-engineering-2/agent", "manifestUrl": "https://openagent3.xyz/skills/prompt-engineering-2/agent.json", "briefUrl": "https://openagent3.xyz/skills/prompt-engineering-2/agent.md" }, "agentAssist": { "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.", "steps": [ "Download the package from Yavira.", "Extract it into a folder your agent can access.", "Paste one of the prompts below and point your agent at the extracted folder." ], "prompts": [ { "label": "New install", "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete." }, { "label": "Upgrade existing", "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run." } ] }, "documentation": { "source": "clawhub", "primaryDoc": "SKILL.md", "sections": [ { "title": "Overview", "body": "This skill transforms vague user requests into precise, effective prompts through collaborative dialogue, systematic analysis, and iterative refinement. It combines proven prompt engineering techniques with a structured development process to create prompts that reliably achieve user objectives." }, { "title": "Workflow Decision Tree", "body": "When a user requests prompt assistance, follow this decision flow:\n\nUser Request\n├─ \"Create a prompt\" / \"Make a prompt\" / Vague request\n│ └─ → Start with EXPLORATION PHASE\n├─ \"Optimize this prompt\" / Has existing prompt\n│ └─ → Start with SIMPLE OPTIMIZATION\n└─ \"Fix this issue with my prompt\" / Specific problem\n └─ → Start with ANALYSIS PHASE (focused on problem)" }, { "title": "Phase 1: Exploration - Uncovering True Needs", "body": "Before creating any prompt, deeply understand the user's actual needs through strategic questioning. Start broad, then narrow down systematically.\n\nInitial Context Gathering:\n\nWhat task will this prompt accomplish?\nWho will use it and in what environment?\nHow frequently will it be used?\nWhat does success look like?\n\nDeepening Understanding:\n\nRequest concrete examples of desired outputs\nAsk about past failures or attempts\nIdentify critical success factors\nUncover unstated assumptions and constraints\n\nTechnical Requirements:\n\nModel and platform constraints\nToken limits and cost considerations\nResponse time requirements\nIntegration with other systems\n\nContinue exploration until the core requirements are crystal clear. Never assume—always verify." }, { "title": "Phase 2: Analysis - Choosing the Right Strategy", "body": "Analyze the task to determine the optimal prompting approach.\n\nTask Classification:\n\nClassify the task along key dimensions:\n\nComplexity: Simple directive vs multi-step reasoning\nOutput Type: Creative vs analytical vs structured\nError Tolerance: High-stakes vs experimental\nFrequency: One-time vs repeated use\n\nStrategy Selection:\n\nBased on classification, choose primary techniques:\n\nSimple Tasks: Direct instructions with clear constraints\nComplex Reasoning: Chain-of-thought with step-by-step breakdown\nCreative Tasks: Role setting with flexible boundaries\nStructured Output: Explicit format specifications with examples\nHigh-Stakes: Self-consistency checks and validation steps\n\nTrade-off Analysis:\n\nPresent multiple approaches with clear trade-offs:\n\nApproach A: Detailed but token-heavy\nApproach B: Concise but requires interpretation\nApproach C: Balanced with moderate complexity\n\nAlways explain WHY each approach fits the specific context." }, { "title": "Phase 3: Implementation - Building Iteratively", "body": "Create the prompt through progressive refinement, starting simple and adding complexity as needed.\n\nVersion 1 - Minimal Viable Prompt:\n\nCore instructions only\nTest basic functionality\nIdentify gaps and ambiguities\n\nVersion 2 - Enhanced Clarity:\n\nAdd specific examples if needed\nClarify ambiguous points\nInclude essential constraints\n\nVersion 3+ - Optimization:\n\nRefine wording for precision\nRemove redundancy\nBalance detail with conciseness\n\nDocument each version's changes and rationale. Store prompts in markdown files with:\n\nVersion history\nDesign decisions\nKnown limitations\nUsage examples" }, { "title": "Phase 4: Validation - Critical Evaluation", "body": "Rigorously evaluate the prompt against quality criteria.\n\nEssential Checks:\n\nClarity: Can the instructions be misunderstood?\nCompleteness: Are all necessary elements present?\nConsistency: Do instructions contradict each other?\nEfficiency: Can anything be removed without loss?\nRobustness: How does it handle edge cases?\n\nTesting Approach:\n\nRun through typical use cases\nTest boundary conditions\nImagine failure modes\nCheck for unwanted behaviors\n\nBe ruthlessly honest about weaknesses. If something isn't working, acknowledge it and iterate." }, { "title": "Simple Optimization", "body": "When optimizing an existing prompt, focus on minimal, targeted improvements:\n\nIdentify Specific Issues: What exactly isn't working?\nDiagnose Root Causes: Why is the current prompt failing?\nApply Minimal Edits: Change only what's necessary\nPreserve Working Elements: Keep what already works well\nTest Improvements: Verify fixes don't break other aspects\n\nCommon optimization targets:\n\nAmbiguous language → Specific instructions\nMissing constraints → Added boundaries\nInconsistent outputs → Format specifications\nVerbose responses → Length constraints\nOff-topic responses → Clearer scope definition" }, { "title": "Prompt Creation from Scratch", "body": "When creating new prompts, structure them as instructions for an eager but inexperienced assistant who needs clear guidance.\n\nEssential Components:\n\nRole/Context (if beneficial):\n\nSet perspective or expertise level\nEstablish tone and approach\n\n\n\nClear Objective:\n\nState the primary goal explicitly\nDefine success criteria\n\n\n\nSpecific Instructions:\n\nBreak complex tasks into steps\nProvide decision criteria\nSpecify constraints and boundaries\n\n\n\nOutput Format (when relevant):\n\nDefine structure explicitly\nProvide format examples\nSpecify length or detail level\n\n\n\nExamples (when clarifying):\n\nShow desired patterns\nIllustrate edge cases\nDemonstrate style/tone" }, { "title": "Foundation Techniques", "body": "Role Setting: Establish perspective when expertise or tone matters\n\nEffective for: Specialized knowledge, consistent voice\nExample: \"As an experienced code reviewer, analyze...\"\n\nProgressive Disclosure: Start general, add detail as needed\n\nEffective for: Complex multi-part tasks\nExample: \"First outline the approach, then implement each section...\"\n\nExplicit Constraints: Define boundaries clearly\n\nEffective for: Preventing unwanted outputs\nExample: \"Limit response to 3 paragraphs, focus only on technical aspects\"" }, { "title": "Advanced Techniques", "body": "Chain-of-Thought: Request reasoning before conclusions\n\nUse when: Logic and transparency matter\nTrigger: \"Think step-by-step\" or \"Explain your reasoning\"\n\nFew-Shot Learning: Provide input-output examples\n\nUse when: Pattern is easier shown than explained\nCaution: 2-3 examples usually sufficient\n\nSelf-Consistency: Have model verify its own outputs\n\nUse when: Accuracy is critical\nImplementation: \"Review your answer for errors and inconsistencies\"\n\nFor detailed technique explanations and examples, consult:\n\nreferences/techniques.md - Comprehensive technique catalog\nreferences/patterns.md - Common prompt patterns\nreferences/antipatterns.md - What to avoid" }, { "title": "Be a Thought Partner, Not Just an Executor", "body": "Bad: \"Here's your prompt\" (without understanding needs)\nGood: \"Let me understand what you're trying to achieve first...\"" }, { "title": "Question Assumptions Constructively", "body": "Surface hidden requirements through dialogue\nChallenge unclear objectives respectfully\nPropose alternatives when original approach seems suboptimal" }, { "title": "Iterate Based on Feedback", "body": "Start with minimum viable prompt\nTest and refine based on actual outputs\nDocument what works and what doesn't" }, { "title": "Teach While Doing", "body": "Explain why certain techniques work\nShare the reasoning behind design choices\nHelp users understand prompt engineering principles" }, { "title": "References", "body": "This skill includes detailed reference documentation:" }, { "title": "references/", "body": "techniques.md - Complete catalog of prompting techniques with examples\npatterns.md - Reusable prompt patterns for common scenarios\nantipatterns.md - Common mistakes and how to avoid them\nevaluation.md - Comprehensive quality evaluation framework\nexamples.md - Library of before/after prompt improvements\n\nConsult these references for in-depth technical details and extensive examples not included in this overview." } ], "body": "Prompt Engineering\nOverview\n\nThis skill transforms vague user requests into precise, effective prompts through collaborative dialogue, systematic analysis, and iterative refinement. It combines proven prompt engineering techniques with a structured development process to create prompts that reliably achieve user objectives.\n\nWorkflow Decision Tree\n\nWhen a user requests prompt assistance, follow this decision flow:\n\nUser Request\n├─ \"Create a prompt\" / \"Make a prompt\" / Vague request\n│ └─ → Start with EXPLORATION PHASE\n├─ \"Optimize this prompt\" / Has existing prompt\n│ └─ → Start with SIMPLE OPTIMIZATION\n└─ \"Fix this issue with my prompt\" / Specific problem\n └─ → Start with ANALYSIS PHASE (focused on problem)\n\nCore Process\nPhase 1: Exploration - Uncovering True Needs\n\nBefore creating any prompt, deeply understand the user's actual needs through strategic questioning. Start broad, then narrow down systematically.\n\nInitial Context Gathering:\n\nWhat task will this prompt accomplish?\nWho will use it and in what environment?\nHow frequently will it be used?\nWhat does success look like?\n\nDeepening Understanding:\n\nRequest concrete examples of desired outputs\nAsk about past failures or attempts\nIdentify critical success factors\nUncover unstated assumptions and constraints\n\nTechnical Requirements:\n\nModel and platform constraints\nToken limits and cost considerations\nResponse time requirements\nIntegration with other systems\n\nContinue exploration until the core requirements are crystal clear. Never assume—always verify.\n\nPhase 2: Analysis - Choosing the Right Strategy\n\nAnalyze the task to determine the optimal prompting approach.\n\nTask Classification:\n\nClassify the task along key dimensions:\n\nComplexity: Simple directive vs multi-step reasoning\nOutput Type: Creative vs analytical vs structured\nError Tolerance: High-stakes vs experimental\nFrequency: One-time vs repeated use\n\nStrategy Selection:\n\nBased on classification, choose primary techniques:\n\nSimple Tasks: Direct instructions with clear constraints\nComplex Reasoning: Chain-of-thought with step-by-step breakdown\nCreative Tasks: Role setting with flexible boundaries\nStructured Output: Explicit format specifications with examples\nHigh-Stakes: Self-consistency checks and validation steps\n\nTrade-off Analysis:\n\nPresent multiple approaches with clear trade-offs:\n\nApproach A: Detailed but token-heavy\nApproach B: Concise but requires interpretation\nApproach C: Balanced with moderate complexity\n\nAlways explain WHY each approach fits the specific context.\n\nPhase 3: Implementation - Building Iteratively\n\nCreate the prompt through progressive refinement, starting simple and adding complexity as needed.\n\nVersion 1 - Minimal Viable Prompt:\n\nCore instructions only\nTest basic functionality\nIdentify gaps and ambiguities\n\nVersion 2 - Enhanced Clarity:\n\nAdd specific examples if needed\nClarify ambiguous points\nInclude essential constraints\n\nVersion 3+ - Optimization:\n\nRefine wording for precision\nRemove redundancy\nBalance detail with conciseness\n\nDocument each version's changes and rationale. Store prompts in markdown files with:\n\nVersion history\nDesign decisions\nKnown limitations\nUsage examples\nPhase 4: Validation - Critical Evaluation\n\nRigorously evaluate the prompt against quality criteria.\n\nEssential Checks:\n\nClarity: Can the instructions be misunderstood?\nCompleteness: Are all necessary elements present?\nConsistency: Do instructions contradict each other?\nEfficiency: Can anything be removed without loss?\nRobustness: How does it handle edge cases?\n\nTesting Approach:\n\nRun through typical use cases\nTest boundary conditions\nImagine failure modes\nCheck for unwanted behaviors\n\nBe ruthlessly honest about weaknesses. If something isn't working, acknowledge it and iterate.\n\nSimple Optimization\n\nWhen optimizing an existing prompt, focus on minimal, targeted improvements:\n\nIdentify Specific Issues: What exactly isn't working?\nDiagnose Root Causes: Why is the current prompt failing?\nApply Minimal Edits: Change only what's necessary\nPreserve Working Elements: Keep what already works well\nTest Improvements: Verify fixes don't break other aspects\n\nCommon optimization targets:\n\nAmbiguous language → Specific instructions\nMissing constraints → Added boundaries\nInconsistent outputs → Format specifications\nVerbose responses → Length constraints\nOff-topic responses → Clearer scope definition\nPrompt Creation from Scratch\n\nWhen creating new prompts, structure them as instructions for an eager but inexperienced assistant who needs clear guidance.\n\nEssential Components:\n\nRole/Context (if beneficial):\n\nSet perspective or expertise level\nEstablish tone and approach\n\nClear Objective:\n\nState the primary goal explicitly\nDefine success criteria\n\nSpecific Instructions:\n\nBreak complex tasks into steps\nProvide decision criteria\nSpecify constraints and boundaries\n\nOutput Format (when relevant):\n\nDefine structure explicitly\nProvide format examples\nSpecify length or detail level\n\nExamples (when clarifying):\n\nShow desired patterns\nIllustrate edge cases\nDemonstrate style/tone\nKey Techniques Reference\nFoundation Techniques\n\nRole Setting: Establish perspective when expertise or tone matters\n\nEffective for: Specialized knowledge, consistent voice\nExample: \"As an experienced code reviewer, analyze...\"\n\nProgressive Disclosure: Start general, add detail as needed\n\nEffective for: Complex multi-part tasks\nExample: \"First outline the approach, then implement each section...\"\n\nExplicit Constraints: Define boundaries clearly\n\nEffective for: Preventing unwanted outputs\nExample: \"Limit response to 3 paragraphs, focus only on technical aspects\"\nAdvanced Techniques\n\nChain-of-Thought: Request reasoning before conclusions\n\nUse when: Logic and transparency matter\nTrigger: \"Think step-by-step\" or \"Explain your reasoning\"\n\nFew-Shot Learning: Provide input-output examples\n\nUse when: Pattern is easier shown than explained\nCaution: 2-3 examples usually sufficient\n\nSelf-Consistency: Have model verify its own outputs\n\nUse when: Accuracy is critical\nImplementation: \"Review your answer for errors and inconsistencies\"\n\nFor detailed technique explanations and examples, consult:\n\nreferences/techniques.md - Comprehensive technique catalog\nreferences/patterns.md - Common prompt patterns\nreferences/antipatterns.md - What to avoid\nCollaboration Principles\nBe a Thought Partner, Not Just an Executor\nBad: \"Here's your prompt\" (without understanding needs)\nGood: \"Let me understand what you're trying to achieve first...\"\nQuestion Assumptions Constructively\nSurface hidden requirements through dialogue\nChallenge unclear objectives respectfully\nPropose alternatives when original approach seems suboptimal\nIterate Based on Feedback\nStart with minimum viable prompt\nTest and refine based on actual outputs\nDocument what works and what doesn't\nTeach While Doing\nExplain why certain techniques work\nShare the reasoning behind design choices\nHelp users understand prompt engineering principles\nReferences\n\nThis skill includes detailed reference documentation:\n\nreferences/\ntechniques.md - Complete catalog of prompting techniques with examples\npatterns.md - Reusable prompt patterns for common scenarios\nantipatterns.md - Common mistakes and how to avoid them\nevaluation.md - Comprehensive quality evaluation framework\nexamples.md - Library of before/after prompt improvements\n\nConsult these references for in-depth technical details and extensive examples not included in this overview." }, "trust": { "sourceLabel": "tencent", "provenanceUrl": "https://clawhub.ai/kongyo2/prompt-engineering-2", "publisherUrl": "https://clawhub.ai/kongyo2/prompt-engineering-2", "owner": "kongyo2", "version": "0.1.0", "license": null, "verificationStatus": "Indexed source record" }, "links": { "detailUrl": "https://openagent3.xyz/skills/prompt-engineering-2", "downloadUrl": "https://openagent3.xyz/downloads/prompt-engineering-2", "agentUrl": "https://openagent3.xyz/skills/prompt-engineering-2/agent", "manifestUrl": "https://openagent3.xyz/skills/prompt-engineering-2/agent.json", "briefUrl": "https://openagent3.xyz/skills/prompt-engineering-2/agent.md" } }