{"id":"e43f69a3-2874-457f-b13c-1c55b7baebe7","shortId":"HWfE7d","kind":"skill","title":"prompt-architect","tagline":"Analyzes and improves prompts using 27 research-backed frameworks across 7 intent categories. Use when a user wants to improve, rewrite, structure, or engineer a prompt — including requests like \"help me write a better prompt\", \"improve this prompt\", \"what framework should I use\"","description":"# Prompt Architect\n\nYou are an expert in prompt engineering and systematic application of prompting frameworks. Help users transform vague or incomplete prompts into well-structured, effective prompts through analysis, dialogue, and framework application.\n\n## Core Process\n\n### 1. Initial Assessment\n\nWhen a user provides a prompt to improve, analyze across dimensions:\n- **Clarity**: Is the goal clear and unambiguous?\n- **Specificity**: Are requirements detailed enough?\n- **Context**: Is necessary background provided?\n- **Constraints**: Are limitations specified?\n- **Output Format**: Is desired format clear?\n\n### 2. Intent-Based Framework Selection\n\nWith 27 frameworks, identify the user's **primary intent** first, then use the discriminating questions within that category.\n\n---\n\n**A. RECOVER** — Reconstruct a prompt from an existing output\n→ **RPEF** (Reverse Prompt Engineering)\n*Signal: \"I have a good output but need/lost the prompt\"*\n\n---\n\n**B. CLARIFY** — Requirements are unclear; gather information first\n→ **Reverse Role Prompting** (AI-Led Interview)\n*Signal: \"I know roughly what I want but struggle to specify the details\"*\n\n---\n\n**C. CREATE** — Generating new content from scratch\n\n| Signal | Framework |\n|--------|-----------|\n| Ultra-minimal, one-off | **APE** |\n| Simple, expertise-driven | **RTF** |\n| Simple, context/situation-driven | **CTF** |\n| Role + context + explicit outcome needed | **RACE** |\n| Multiple output variants needed | **CRISPE** |\n| Business deliverable with KPIs | **BROKE** |\n| Explicit rules/compliance constraints | **CARE** or **TIDD-EC** |\n| Audience, tone, style are critical | **CO-STAR** |\n| Multi-step procedure or methodology | **RISEN** |\n| Data transformation (input → output) | **RISE-IE** |\n| Content creation with reference examples | **RISE-IX** |\n\n*TIDD-EC vs. CARE: separate Do/Don't lists → TIDD-EC; combined rules + examples → CARE*\n\n---\n\n**D. TRANSFORM** — Improving or converting existing content\n\n| Signal | Framework |\n|--------|-----------|\n| Rewrite, refactor, convert | **BAB** |\n| Iterative quality improvement | **Self-Refine** |\n| Compress or densify | **Chain of Density** |\n| Outline-first then expand sections | **Skeleton of Thought** |\n\n---\n\n**E. REASON** — Solving a reasoning or calculation problem\n\n| Signal | Framework |\n|--------|-----------|\n| Numerical/calculation, zero-shot | **Plan-and-Solve (PS+)** |\n| Multi-hop with ordered dependencies | **Least-to-Most** |\n| Needs first-principles before answering | **Step-Back** |\n| Multiple distinct approaches to compare | **Tree of Thought** |\n| Verify reasoning didn't overlook conditions | **RCoT** |\n| Linear step-by-step reasoning | **Chain of Thought** |\n\n---\n\n**F. CRITIQUE** — Stress-testing, attacking, or verifying output\n\n| Signal | Framework |\n|--------|-----------|\n| General quality improvement | **Self-Refine** |\n| Align to explicit principle/standard | **CAI Critique-Revise** |\n| Find the strongest opposing argument | **Devil's Advocate** |\n| Identify failure modes before they happen | **Pre-Mortem** |\n| Verify reasoning didn't miss conditions | **RCoT** |\n\n*Self-Refine = any quality. CAI = principle compliance. Devil's Advocate = opposing arguments. Pre-Mortem = failure analysis. RCoT = condition verification.*\n\n---\n\n**G. AGENTIC** — Tool-use with iterative reasoning\n→ **ReAct** (Reasoning + Acting)\n*Signal: \"Task requires tools; each result informs the next step\"*\n\n---\n\n### 3. Framework Quick Reference\n\nOne-line per framework (load `references/frameworks/` for full detail):\n\n**Simple:** APE | RTF | CTF\n**Medium:** RACE | CARE | BAB | BROKE | CRISPE\n**Comprehensive:** CO-STAR | RISEN | TIDD-EC\n**Data:** RISE-IE | RISE-IX\n**Reasoning:** Plan-and-Solve | Chain of Thought | Least-to-Most | Step-Back | Tree of Thought | RCoT\n**Structure/Iteration:** Skeleton of Thought | Chain of Density\n**Critique/Quality:** Self-Refine | CAI Critique-Revise | Devil's Advocate | Pre-Mortem\n**Meta/Reverse:** RPEF | Reverse Role Prompting\n**Agentic:** ReAct\n\n### 4. Clarification Questions\n\nAsk targeted questions (3-5 at a time) based on identified gaps:\n\n**For CO-STAR**: Context, audience, tone, style, objective, format?\n**For RISEN**: Role, principles, steps, success criteria, constraints?\n**For RISE-IE**: Role, input format/characteristics, processing steps, output expectations?\n**For RISE-IX**: Role, task instructions, workflow steps, reference examples?\n**For TIDD-EC**: Task type, exact steps, what to include (dos), what to avoid (don'ts), examples, context?\n**For CTF**: What is the situation/background, exact task, output format?\n**For RTF**: Expertise needed, exact task, output format?\n**For APE**: Core action, why it's needed, what success looks like?\n**For BAB**: What is the current state/problem, what should it become, transformation rules?\n**For RACE**: Role/expertise, action, situational context, explicit expectation?\n**For CRISPE**: Capacity/role, background insight, instructions, personality/style, how many variants?\n**For BROKE**: Background situation, role, objective, measurable key results, evolve instructions?\n**For CARE**: Context/situation, specific ask, explicit rules and constraints, examples of good output?\n**For Tree of Thought**: Problem, distinct solution branches to explore, evaluation criteria?\n**For ReAct**: Goal, available tools, constraints and stop condition?\n**For Skeleton of Thought**: Topic/question, number of skeleton points, expansion depth per point?\n**For Step-Back**: Original question, what higher-level principle governs it?\n**For Least-to-Most**: Full problem, decomposed subproblems in dependency order?\n**For Plan-and-Solve**: Problem with all relevant numbers/variables?\n**For Chain of Thought**: Problem, reasoning steps, verification?\n**For Chain of Density**: Content to improve, iterations, optimization goals?\n**For Self-Refine**: Output to improve, feedback dimensions, stop condition?\n**For CAI Critique-Revise**: The principle to enforce, output to critique?\n**For Devil's Advocate**: Position to attack, attack dimensions, severity ranking needed?\n**For Pre-Mortem**: Project/decision, time horizon, domains to analyze?\n**For RCoT**: Question with all conditions, initial answer to verify?\n**For RPEF**: Output sample to reverse-engineer, input data if available?\n**For Reverse Role**: Intent statement, domain of expertise, interview mode (batch vs. conversational)?\n\n### 4. Apply Framework\n\nUsing gathered information:\n1. Load appropriate template from `assets/templates/`\n2. Map user's information to framework components\n3. Fill missing elements with reasonable defaults\n4. Structure according to framework format\n\n### 5. Present Improvements\n\nStructure your output in this exact order:\n\n**A. Analysis section** (comes first):\n- Framework selected and why\n- Changes made and reasoning\n- Framework components applied\n\n**B. Usage instructions** (transition block, immediately before the prompt):\n\n> **Your revised prompt is ready.**\n> - **New chat**: Copy the prompt below and paste it as your first message in a new conversation.\n> - **Same chat**: Tell the assistant: *\"Use the revised prompt you just provided as a new instruction and execute it.\"*\n\n**C. The revised prompt** (comes last, in a fenced code block):\n- Present as a clean, flat-text block inside triple backticks\n- **No framework section headers** (no \"BEFORE:\", \"BRIDGE:\", \"CONTEXT:\", etc.) — these are scaffolding, not part of the deliverable\n- **No indentation** beyond what the prompt itself genuinely requires\n- **No markdown formatting** inside the block unless the prompt explicitly needs it (e.g., it asks for tables)\n- The user must be able to copy the entire block contents and paste it verbatim with zero editing\n- **Nothing after the code block** — the revised prompt must be the absolute last element in the response. No trailing suggestions, tips, or follow-up text after the closing backticks.\n\n### 6. Iterate\n\n- Confirm improvements align with intent\n- Refine based on feedback\n- Switch or combine frameworks if needed\n- Continue until satisfactory\n\n## Framework References\n\nDetailed framework docs in `references/frameworks/`:\n- `co-star.md` - Context, Objective, Style, Tone, Audience, Response\n- `risen.md` - Role, Instructions, Steps, End goal, Narrowing\n- `rise.md` - **Dual variant support**: RISE-IE (Input-Expectation) & RISE-IX (Instructions-Examples)\n- `tidd-ec.md` - Task type, Instructions, Do, Don't, Examples, Context\n- `ctf.md` - Context, Task, Format\n- `rtf.md` - Role, Task, Format\n- `ape.md` - Action, Purpose, Expectation (ultra-minimal)\n- `bab.md` - Before, After, Bridge (transformation/rewrite tasks)\n- `race.md` - Role, Action, Context, Expectation (medium complexity)\n- `crispe.md` - Capacity+Role, Insight, Instructions, Personality, Experiment\n- `broke.md` - Background, Role, Objective, Key Results, Evolve\n- `care.md` - Context, Ask, Rules, Examples (constraint-driven)\n- `tree-of-thought.md` - Branching exploration of multiple solution paths\n- `react.md` - Reasoning + Acting (agentic tool-use cycles)\n- `skeleton-of-thought.md` - Skeleton-first then expand (parallel generation)\n- `step-back.md` - Abstract to principles first, then answer (Google DeepMind)\n- `least-to-most.md` - Decompose into ordered subproblems, solve sequentially\n- `plan-and-solve.md` - Zero-shot: plan + extract variables + calculate (PS+)\n- `chain-of-thought.md` - Step-by-step reasoning techniques\n- `chain-of-density.md` - Iterative refinement through compression\n- `self-refine.md` - Generate → Feedback → Refine loop (NeurIPS 2023)\n- `cai-critique-revise.md` - Principle-based critique + revision (Anthropic)\n- `devils-advocate.md` - Strongest opposing argument generation (ACM IUI 2024)\n- `pre-mortem.md` - Assume failure, identify causes + warning signs (Gary Klein)\n- `rcot.md` - Reverse Chain-of-Thought: verify by reconstructing the question\n- `rpef.md` - Reverse Prompt Engineering: recover prompt from output (EMNLP 2025)\n- `reverse-role.md` - AI-Led Interview: AI asks you questions first (FATA)\n\nLoad these when applying specific frameworks for detailed component guidance, selection criteria, and examples.\n\n## Templates\n\nFramework templates in `assets/templates/` provide structure:\n- `co-star_template.txt` - Full CO-STAR structure\n- `risen_template.txt` - Full RISEN structure\n- `rise-ie_template.txt` - RISE-IE structure (Input-Expectation for data tasks)\n- `rise-ix_template.txt` - RISE-IX structure (Instructions-Examples for creative tasks)\n- `tidd-ec_template.txt` - TIDD-EC structure (Task, Instructions, Do, Don't, Examples, Context)\n- `ctf_template.txt` - CTF structure (Context-Task-Format for situational prompts)\n- `rtf_template.txt` - Full RTF structure\n- `ape_template.txt` - APE structure (Action-Purpose-Expectation ultra-minimal)\n- `bab_template.txt` - BAB structure (Before-After-Bridge for transformations)\n- `race_template.txt` - RACE structure (Role-Action-Context-Expectation)\n- `crispe_template.txt` - CRISPE structure (with Experiment/variants)\n- `broke_template.txt` - BROKE structure (with Key Results + Evolve)\n- `care_template.txt` - CARE structure (with Rules + Examples)\n- `tree-of-thought_template.txt` - Tree of Thought branching exploration structure\n- `react_template.txt` - ReAct Thought-Action-Observation cycle structure\n- `skeleton-of-thought_template.txt` - Skeleton + expand structure\n- `step-back_template.txt` - Step-back question + principle application\n- `least-to-most_template.txt` - Decompose + sequential solving\n- `plan-and-solve_template.txt` - PS+ trigger phrase structure\n- `chain-of-thought_template.txt` - Step-by-step reasoning with verification\n- `chain-of-density_template.txt` - Iterative compression with stopping criterion\n- `self-refine_template.txt` - Generate → Feedback → Refine structure\n- `cai-critique-revise_template.txt` - Principle → Critique → Revision structure\n- `devils-advocate_template.txt` - Position attack with severity ranking\n- `pre-mortem_template.txt` - Failure assumption + cause analysis\n- `rcot_template.txt` - 4-step backward verification structure\n- `rpef_template.txt` - Output analysis + recovered prompt template\n- `reverse-role_template.txt` - Intent + interview trigger structure\n- `hybrid_template.txt` - Combined framework approach\n\n## Key Principles\n\n1. **Ask Before Assuming** - Don't guess intent; clarify ambiguities\n2. **Explain Reasoning** - Why this framework? Why these changes?\n3. **Show Your Work** - Display analysis, show framework mapping\n4. **Be Iterative** - Start with analysis, refine progressively\n5. **Respect User Choices** - Adapt if user prefers different framework\n\n## When NOT to Use Frameworks\n\nFrameworks add structure — but structure has overhead. Skip them when:\n\n- **The prompt is already complete**: Clear goal, full context, defined format → just execute it.\n- **Purely factual lookups**: \"What is the capital of France?\" — no framework needed.\n- **Conversational exchanges**: Back-and-forth dialogue doesn't need a structured template.\n- **Very short one-off tasks**: \"Translate this sentence to Spanish.\" APE would be overhead; just translate.\n- **User is in a hurry**: If someone explicitly says \"just do it\", don't pause for framework selection — deliver, then offer to structure if they want more.\n- **The task is fully specced by context**: When the codebase, existing docs, or prior messages already contain everything needed.\n\n**Rule of thumb**: Apply a framework when there's a gap between what the user *asked for* and what they *need*. If there's no gap, there's no job for a framework.\n\n---\n\n## Example Interaction\n\nThis example shows the full intent-based flow — detecting intent, using the discriminating table, selecting a framework, asking questions, and delivering a structured result.\n\n**User**: \"Rewrite this job posting — it's too formal and isn't getting applicants.\"\n\n**Assistant**: I'll analyze this and identify the best approach.\n\n**Intent detected**: TRANSFORM — improving existing content.\n\n**Discriminating question**: Is this iterative quality improvement, or a one-shot before-to-after rewrite?\n→ One-shot rewrite with a clear current state → **BAB** (Before, After, Bridge)\n\n**Questions**:\n1. What's the role and target audience? (who should this attract?)\n2. What does \"too formal\" mean specifically? (stiff language, corporate jargon, passive voice?)\n3. What tone should the new version have? (casual-professional, startup-energy, warm?)\n4. Any constraints to preserve? (job requirements, company name, legal language?)\n5. How much can change? (light edits vs. full rewrite?)\n\n**User**: \"Software engineer, early-career devs. Too much corporate-speak. Want it to sound like real humans work there. Requirements must stay. Full rewrite OK.\"\n\n**Analysis (BAB framework applied)**:\n1. Locked the current state so the AI understands the starting point\n2. Defined the target state in terms the AI can evaluate against\n3. Made transformation rules explicit and prioritized\n4. Protected non-negotiable elements (requirements) from being changed\n5. Gave a concrete length/tone constraint to prevent over-engineering\n\n> **Your revised prompt is ready.**\n> - **New chat**: Copy the prompt below and paste it as your first message in a new conversation.\n> - **Same chat**: Tell the assistant: *\"Use the revised prompt you just provided as a new instruction and execute it.\"*\n\n```\nRewrite the following job posting. The current version suffers from corporate-speak, passive voice, overly formal tone, and generic language that doesn't reflect actual team culture.\n\n[Paste the current job posting here]\n\nThe rewritten version should sound like it was written by engineers, for engineers. Early-career developers should read it and think \"I want to work there.\" It should feel honest, direct, and human — not like legal boilerplate.\n\nFollow these rules:\n- Replace all passive constructions with active voice.\n- Convert corporate jargon to plain English (e.g., \"leverage\" → \"use\").\n- Add one specific, concrete detail about the team or culture per section.\n- Keep all technical requirements and must-haves verbatim — do not change these.\n- Target reading level: conversational, not academic.\n- Length: same or shorter than the original. Cut fluff, don't add it.\n```\n\n---\n\n## Usage Notes\n\n- Always start by analyzing the original prompt\n- Recommend framework(s) with reasoning\n- Ask clarifying questions progressively (don't overwhelm)\n- Apply framework systematically using templates\n- Present improvements with explanation\n- Iterate based on feedback\n- Load framework references only when needed for detailed guidance","tags":["prompt","architect","ckelsoe","agent-skills","ai-coding-agents","ai-tools","chain-of-thought","chatgpt","claude-code","claude-skill","cursor","gemini-cli"],"capabilities":["skill","source-ckelsoe","skill-prompt-architect","topic-agent-skills","topic-ai-coding-agents","topic-ai-tools","topic-chain-of-thought","topic-chatgpt","topic-claude-code","topic-claude-skill","topic-cursor","topic-gemini-cli","topic-github-copilot","topic-llm-prompts","topic-openai-codex"],"categories":["prompt-architect"],"synonyms":[],"warnings":[],"endpointUrl":"https://skills.sh/ckelsoe/prompt-architect/prompt-architect","protocol":"skill","transport":"skills-sh","auth":{"type":"none","details":{"cli":"npx skills add ckelsoe/prompt-architect","source_repo":"https://github.com/ckelsoe/prompt-architect","install_from":"skills.sh"}},"qualityScore":"0.522","qualityRationale":"deterministic score 0.52 from registry signals: · indexed on github topic:agent-skills · 144 github stars · SKILL.md body (16,604 chars)","verified":false,"liveness":"unknown","lastLivenessCheck":null,"agentReviews":{"count":0,"score_avg":null,"cost_usd_avg":null,"success_rate":null,"latency_p50_ms":null,"narrative_summary":null,"summary_updated_at":null},"enrichmentModel":"deterministic:skill-github:v1","enrichmentVersion":1,"enrichedAt":"2026-05-02T12:54:45.789Z","embedding":null,"createdAt":"2026-04-18T22:10:28.566Z","updatedAt":"2026-05-02T12:54:45.789Z","lastSeenAt":"2026-05-02T12:54:45.789Z","tsv":"'-5':574 '1':84,899,1571,1855,1947 '2':125,905,1581,1867,1959 '2023':1297 '2024':1312 '2025':1342 '27':9,132 '3':481,573,913,1590,1880,1971 '4':567,893,920,1549,1599,1895,1978 '5':926,1607,1906,1988 '6':1115 '7':15 'abl':1071 'absolut':1096 'abstract':1255 'academ':2161 'accord':922 'acm':1310 'across':14,96 'act':470,1240 'action':662,687,1190,1204,1437,1457,1489 'action-purpose-expect':1436 'activ':2120 'actual':2065 'adapt':1611 'add':1623,2131,2173 'advoc':422,449,556,839 'agent':461,565,1241 'ai':184,1345,1348,1954,1967 'ai-l':183,1344 'align':407,1119 'alreadi':1635,1730 'alway':2177 'ambigu':1580 'analysi':77,456,937,1547,1556,1595,1604,1943 'analyz':4,95,857,1811,2180 'answer':362,865,1260 'anthrop':1304 'ape':215,496,660,1434,1682 'ape.md':1189 'ape_template.txt':1433 'appli':894,951,1357,1737,1946,2196 'applic':59,81,1503,1807 'approach':368,1568,1817 'appropri':901 'architect':3,49 'argument':419,451,1308 'ask':570,717,1064,1225,1349,1572,1749,1787,2189 'assess':86 'assets/templates':904,1372 'assist':987,1808,2025 'assum':1314,1574 'assumpt':1545 'attack':395,842,843,1539 'attract':1866 'audienc':248,587,1147,1862 'avail':741,879 'avoid':636 'b':172,952 'bab':306,502,672,1444,1850,1944 'bab.md':1196 'bab_template.txt':1443 'back':12,365,534,763,1500,1661 'back-and-forth':1660 'background':113,695,704,1217 'backtick':1023,1114 'backward':1551 'base':128,578,1123,1301,1776,2206 'batch':890 'becom':681 'before-after-bridg':1446 'before-to-aft':1836 'best':1816 'better':38 'beyond':1043 'block':956,1012,1020,1055,1076,1089 'boilerpl':2111 'branch':733,1232,1482 'bridg':1030,1199,1449,1853 'broke':239,503,703,1466 'broke.md':1216 'broke_template.txt':1465 'busi':235 'c':200,1002 'cai':411,444,550,825 'cai-critique-revise.md':1298 'cai-critique-revise_template.txt':1532 'calcul':334,1277 'capac':1210 'capacity/role':694 'capit':1652 'care':243,282,293,501,714,1473 'care.md':1223 'care_template.txt':1472 'career':1921,2089 'casual':1889 'casual-profession':1888 'categori':17,148 'caus':1317,1546 'chain':316,387,525,543,796,804,1325 'chain-of-density.md':1286 'chain-of-density_template.txt':1521 'chain-of-thought':1324 'chain-of-thought.md':1279 'chain-of-thought_template.txt':1513 'chang':945,1589,1910,1987,2154 'chat':967,984,2005,2022 'choic':1610 'clarif':568 'clarifi':173,1579,2190 'clariti':98 'clean':1016 'clear':102,124,1637,1847 'close':1113 'co':254,507,584,1378 'co-star':253,506,583,1377 'co-star.md':1142 'co-star_template.txt':1375 'code':1011,1088 'codebas':1724 'combin':290,1128,1566 'come':939,1006 'compani':1902 'compar':370 'complet':1636 'complex':1208 'complianc':446 'compon':912,950,1362 'comprehens':505 'compress':313,1290,1523 'concret':1991,2134 'condit':379,437,458,746,823,863 'confirm':1117 'constraint':115,242,599,721,743,1229,1897,1993 'constraint-driven':1228 'construct':2118 'contain':1731 'content':204,270,300,807,1077,1823 'context':110,225,586,640,689,1031,1143,1180,1182,1205,1224,1418,1423,1458,1640,1721 'context-task-format':1422 'context/situation':715 'context/situation-driven':222 'continu':1132 'convers':892,982,1658,2020,2159 'convert':298,305,2122 'copi':968,1073,2006 'core':82,661 'corpor':1876,1926,2051,2123 'corporate-speak':1925,2050 'creat':201 'creation':271 'creativ':1405 'crisp':234,504,693,1461 'crispe.md':1209 'crispe_template.txt':1460 'criteria':598,737,1365 'criterion':1526 'critic':252 'critiqu':391,413,552,827,835,1302,1534 'critique-revis':412,551,826 'critique/quality':546 'ctf':223,498,642,1420 'ctf.md':1181 'ctf_template.txt':1419 'cultur':2067,2140 'current':676,1848,1950,2046,2070 'cut':2169 'cycl':1245,1491 'd':294 'data':263,513,877,1394 'decompos':780,1264,1505 'deepmind':1262 'default':919 'defin':1641,1960 'deliv':1706,1790 'deliver':236,1040 'densifi':315 'densiti':318,545,806 'depend':352,783 'depth':757 'desir':122 'detail':108,199,494,1137,1361,2135,2216 'detect':1778,1819 'dev':1922 'develop':2090 'devil':420,447,554,837 'devils-advocate.md':1305 'devils-advocate_template.txt':1537 'dialogu':78,1664 'didn':376,434 'differ':1615 'dimens':97,821,844 'direct':2105 'discrimin':144,1782,1824 'display':1594 'distinct':367,731 'do/don':284 'doc':1139,1726 'doesn':1665,2062 'domain':855,885 'dos':633 'driven':219,1230 'dual':1157 'e':328 'e.g':1062,2128 'earli':1920,2088 'early-car':1919,2087 'ec':247,280,289,512,625,1410 'edit':1084,1912 'effect':74 'element':916,1098,1983 'emnlp':1341 'end':1153 'energi':1893 'enforc':832 'engin':28,56,161,875,1336,1918,1998,2084,2086 'english':2127 'enough':109 'entir':1075 'etc':1032 'evalu':736,1969 'everyth':1732 'evolv':711,1222,1471 'exact':628,647,655,934 'exampl':274,292,621,639,722,1171,1179,1227,1367,1403,1417,1477,1767,1770 'exchang':1659 'execut':1000,1644,2038 'exist':156,299,1725,1822 'expand':323,1251,1495 'expans':756 'expect':610,691,1165,1192,1206,1392,1439,1459 'experi':1215 'experiment/variants':1464 'expert':53 'expertis':218,653,887 'expertise-driven':217 'explain':1582 'explan':2204 'explicit':226,240,409,690,718,1059,1695,1975 'explor':735,1233,1483 'extract':1275 'f':390 'factual':1647 'failur':424,455,1315,1544 'fata':1353 'feedback':820,1125,1293,1529,2208 'feel':2103 'fenc':1010 'fill':914 'find':415 'first':140,179,321,359,940,977,1249,1258,1352,2015 'first-principl':358 'flat':1018 'flat-text':1017 'flow':1777 'fluff':2170 'follow':1108,2042,2112 'follow-up':1107 'formal':1802,1871,2056 'format':120,123,591,650,658,925,1052,1184,1188,1425,1642 'format/characteristics':606 'forth':1663 'framework':13,44,62,80,129,133,208,302,337,400,482,489,895,911,924,941,949,1025,1129,1135,1138,1359,1369,1567,1586,1597,1616,1621,1622,1656,1704,1739,1766,1786,1945,2185,2197,2210 'franc':1654 'full':493,778,1376,1382,1430,1639,1773,1914,1940 'fulli':1718 'g':460 'gap':581,1744,1759 'gari':1320 'gather':177,897 'gave':1989 'general':401 'generat':202,1253,1292,1309,1528 'generic':2059 'genuin':1048 'get':1806 'goal':101,740,812,1154,1638 'good':166,724 'googl':1261 'govern':771 'guess':1577 'guidanc':1363,2217 'happen':428 'have':2150 'header':1027 'help':34,63 'higher':768 'higher-level':767 'honest':2104 'hop':349 'horizon':854 'human':1934,2107 'hurri':1692 'hybrid_template.txt':1565 'identifi':134,423,580,1316,1814 'ie':269,516,603,1162,1388 'immedi':957 'improv':6,24,40,94,296,309,403,809,819,928,1118,1821,1830,2202 'includ':31,632 'incomplet':68 'indent':1042 'inform':178,477,898,909 'initi':85,864 'input':265,605,876,1164,1391 'input-expect':1163,1390 'insid':1021,1053 'insight':696,1212 'instruct':617,697,712,954,998,1151,1170,1175,1213,1402,1413,2036 'instructions-exampl':1169,1401 'intent':16,127,139,883,1121,1561,1578,1775,1779,1818 'intent-bas':126,1774 'interact':1768 'interview':186,888,1347,1562 'isn':1804 'iter':307,466,810,1116,1287,1522,1601,1828,2205 'iui':1311 'ix':277,519,614,1168,1399 'jargon':1877,2124 'job':1763,1797,1900,2043,2071 'keep':2143 'key':709,1220,1469,1569 'klein':1321 'know':189 'kpis':238 'languag':1875,1905,2060 'last':1007,1097 'least':354,529,775 'least-to-most':353,528,774 'least-to-most.md':1263 'least-to-most_template.txt':1504 'led':185,1346 'legal':1904,2110 'length':2162 'length/tone':1992 'level':769,2158 'leverag':2129 'light':1911 'like':33,670,1932,2079,2109 'limit':117 'line':487 'linear':381 'list':286 'll':1810 'load':490,900,1354,2209 'lock':1948 'look':669 'lookup':1648 'loop':1295 'made':946,1972 'mani':700 'map':906,1598 'markdown':1051 'mean':1872 'measur':708 'medium':499,1207 'messag':978,1729,2016 'meta/reverse':560 'methodolog':261 'minim':211,1195,1442 'miss':436,915 'mode':425,889 'mortem':431,454,559,851 'much':1908,1924 'multi':257,348 'multi-hop':347 'multi-step':256 'multipl':230,366,1235 'must':1069,1093,1938,2149 'must-hav':2148 'name':1903 'narrow':1155 'necessari':112 'need':228,233,357,654,666,847,1060,1131,1657,1667,1733,1754,2214 'need/lost':169 'negoti':1982 'neurip':1296 'new':203,966,981,997,1885,2004,2019,2035 'next':479 'non':1981 'non-negoti':1980 'note':2176 'noth':1085 'number':752 'numbers/variables':794 'numerical/calculation':338 'object':590,707,1144,1219 'observ':1490 'offer':1708 'ok':1942 'one':213,486,1674,1834,1842,2132 'one-lin':485 'one-off':212,1673 'one-shot':1833,1841 'oppos':418,450,1307 'optim':811 'order':351,784,935,1266 'origin':764,2168,2182 'outcom':227 'outlin':320 'outline-first':319 'output':119,157,167,231,266,398,609,649,657,725,817,833,870,931,1340,1555 'over':2055 'over-engin':1996 'overhead':1628,1685 'overlook':378 'overwhelm':2195 'parallel':1252 'part':1037 'passiv':1878,2053,2117 'past':973,1079,2011,2068 'path':1237 'paus':1702 'per':488,758,2141 'person':1214 'personality/style':698 'phrase':1511 'plain':2126 'plan':343,522,787,1274 'plan-and-solv':342,521,786 'plan-and-solve.md':1270 'plan-and-solve_template.txt':1508 'point':755,759,1958 'posit':840,1538 'post':1798,2044,2072 'pre':430,453,558,850 'pre-mortem':429,452,557,849 'pre-mortem.md':1313 'pre-mortem_template.txt':1543 'prefer':1614 'present':927,1013,2201 'preserv':1899 'prevent':1995 'primari':138 'principl':360,445,595,770,830,1257,1300,1502,1533,1570 'principle-bas':1299 'principle/standard':410 'prior':1728 'priorit':1977 'problem':335,730,779,790,799 'procedur':259 'process':83,607 'profession':1890 'progress':1606,2192 'project/decision':852 'prompt':2,7,30,39,42,48,55,61,69,75,92,153,160,171,182,564,960,963,970,991,1005,1046,1058,1092,1335,1338,1428,1558,1633,2001,2008,2029,2183 'prompt-architect':1 'protect':1979 'provid':90,114,994,1373,2032 'ps':346,1278,1509 'pure':1646 'purpos':1191,1438 'qualiti':308,402,443,1829 'question':145,569,572,765,860,1332,1351,1501,1788,1825,1854,2191 'quick':483 'race':229,500,685,1453 'race.md':1202 'race_template.txt':1452 'rank':846,1542 'rcot':380,438,457,538,859 'rcot.md':1322 'rcot_template.txt':1548 'react':468,566,739,1486 'react.md':1238 'react_template.txt':1485 'read':2092,2157 'readi':965,2003 'real':1933 'reason':329,332,375,386,433,467,469,520,800,918,948,1239,1284,1518,1583,2188 'recommend':2184 'reconstruct':151,1330 'recov':150,1337,1557 'refactor':304 'refer':273,484,620,1136,2211 'references/frameworks':491,1141 'refin':312,406,441,549,816,1122,1288,1294,1530,1605 'reflect':2064 'relev':793 'replac':2115 'request':32 'requir':107,174,473,1049,1901,1937,1984,2146 'research':11 'research-back':10 'respect':1608 'respons':1101,1148 'result':476,710,1221,1470,1793 'revers':159,180,562,874,881,1323,1334 'reverse-engin':873 'reverse-role.md':1343 'reverse-role_template.txt':1560 'revis':414,553,828,962,990,1004,1091,1303,1535,2000,2028 'rewrit':25,303,1795,1840,1844,1915,1941,2040 'rewritten':2075 'rise':268,276,515,518,602,613,1161,1167,1387,1398 'rise-i':267,514,601,1160,1386 'rise-ie_template.txt':1385 'rise-ix':275,517,612,1166,1397 'rise-ix_template.txt':1396 'rise.md':1156 'risen':262,509,593,1383 'risen.md':1149 'risen_template.txt':1381 'role':181,224,563,594,604,615,706,882,1150,1186,1203,1211,1218,1456,1859 'role-action-context-expect':1455 'role/expertise':686 'rough':190 'rpef':158,561,869 'rpef.md':1333 'rpef_template.txt':1554 'rtf':220,497,652,1431 'rtf.md':1185 'rtf_template.txt':1429 'rule':291,683,719,1226,1476,1734,1974,2114 'rules/compliance':241 'sampl':871 'satisfactori':1134 'say':1696 'scaffold':1035 'scratch':206 'section':324,938,1026,2142 'select':130,942,1364,1705,1784 'self':311,405,440,548,815 'self-refin':310,404,439,547,814 'self-refine.md':1291 'self-refine_template.txt':1527 'sentenc':1679 'separ':283 'sequenti':1269,1506 'sever':845,1541 'short':1672 'shorter':2165 'shot':341,1273,1835,1843 'show':1591,1596,1771 'sign':1319 'signal':162,187,207,301,336,399,471 'simpl':216,221,495 'situat':688,705,1427 'situation/background':646 'skeleton':325,540,748,754,1248,1494 'skeleton-first':1247 'skeleton-of-thought.md':1246 'skeleton-of-thought_template.txt':1493 'skill' 'skill-prompt-architect' 'skip':1629 'softwar':1917 'solut':732,1236 'solv':330,345,524,789,1268,1507 'someon':1694 'sound':1931,2078 'source-ckelsoe' 'spanish':1681 'speak':1927,2052 'specc':1719 'specif':105,716,1358,1873,2133 'specifi':118,197 'star':255,508,585,1379 'start':1602,1957,2178 'startup':1892 'startup-energi':1891 'state':1849,1951,1963 'state/problem':677 'statement':884 'stay':1939 'step':258,364,383,385,480,533,596,608,619,629,762,801,1152,1281,1283,1499,1515,1517,1550 'step-back':363,532,761,1498 'step-back.md':1254 'step-back_template.txt':1497 'step-by-step':382,1280,1514 'stiff':1874 'stop':745,822,1525 'stress':393 'stress-test':392 'strongest':417,1306 'structur':26,73,921,929,1374,1380,1384,1389,1400,1411,1421,1432,1435,1445,1454,1462,1467,1474,1484,1492,1496,1512,1531,1536,1553,1564,1624,1626,1669,1710,1792 'structure/iteration':539 'struggl':195 'style':250,589,1145 'subproblem':781,1267 'success':597,668 'suffer':2048 'suggest':1104 'support':1159 'switch':1126 'systemat':58,2198 'tabl':1066,1783 'target':571,1861,1962,2156 'task':472,616,626,648,656,1173,1183,1187,1201,1395,1406,1412,1424,1676,1716 'team':2066,2138 'technic':2145 'techniqu':1285 'tell':985,2023 'templat':902,1368,1370,1559,1670,2200 'term':1965 'test':394 'text':1019,1110 'think':2095 'thought':327,373,389,527,537,542,729,750,798,1327,1481,1488 'thought-action-observ':1487 'thumb':1736 'tidd':246,279,288,511,624,1409 'tidd-ec':245,278,287,510,623,1408 'tidd-ec.md':1172 'tidd-ec_template.txt':1407 'time':577,853 'tip':1105 'tone':249,588,1146,1882,2057 'tool':463,474,742,1243 'tool-us':462,1242 'topic-agent-skills' 'topic-ai-coding-agents' 'topic-ai-tools' 'topic-chain-of-thought' 'topic-chatgpt' 'topic-claude-code' 'topic-claude-skill' 'topic-cursor' 'topic-gemini-cli' 'topic-github-copilot' 'topic-llm-prompts' 'topic-openai-codex' 'topic/question':751 'trail':1103 'transform':65,264,295,682,1451,1820,1973 'transformation/rewrite':1200 'transit':955 'translat':1677,1687 'tree':371,535,727,1479 'tree-of-thought.md':1231 'tree-of-thought_template.txt':1478 'trigger':1510,1563 'tripl':1022 'ts':638 'type':627,1174 'ultra':210,1194,1441 'ultra-minim':209,1193,1440 'unambigu':104 'unclear':176 'understand':1955 'unless':1056 'usag':953,2175 'use':8,18,47,142,464,896,988,1244,1620,1780,2026,2130,2199 'user':21,64,89,136,907,1068,1609,1613,1688,1748,1794,1916 'vagu':66 'variabl':1276 'variant':232,701,1158 'verbatim':1081,2151 'verif':459,802,1520,1552 'verifi':374,397,432,867,1328 'version':1886,2047,2076 'voic':1879,2054,2121 'vs':281,891,1913 'want':22,193,1713,1928,2097 'warm':1894 'warn':1318 'well':72 'well-structur':71 'within':146 'work':1593,1935,2099 'workflow':618 'would':1683 'write':36 'written':2082 'zero':340,1083,1272 'zero-shot':339,1271","prices":[{"id":"7325f719-5f2c-4dd2-931e-47d137c5ef2e","listingId":"e43f69a3-2874-457f-b13c-1c55b7baebe7","amountUsd":"0","unit":"free","nativeCurrency":null,"nativeAmount":null,"chain":null,"payTo":null,"paymentMethod":"skill-free","isPrimary":true,"details":{"org":"ckelsoe","category":"prompt-architect","install_from":"skills.sh"},"createdAt":"2026-04-18T22:10:28.566Z"}],"sources":[{"listingId":"e43f69a3-2874-457f-b13c-1c55b7baebe7","source":"github","sourceId":"ckelsoe/prompt-architect/prompt-architect","sourceUrl":"https://github.com/ckelsoe/prompt-architect/tree/main/skills/prompt-architect","isPrimary":false,"firstSeenAt":"2026-04-18T22:10:28.566Z","lastSeenAt":"2026-05-02T12:54:45.789Z"}],"details":{"listingId":"e43f69a3-2874-457f-b13c-1c55b7baebe7","quickStartSnippet":null,"exampleRequest":null,"exampleResponse":null,"schema":null,"openapiUrl":null,"agentsTxtUrl":null,"citations":[],"useCases":[],"bestFor":[],"notFor":[],"kindDetails":{"org":"ckelsoe","slug":"prompt-architect","github":{"repo":"ckelsoe/prompt-architect","stars":144,"topics":["agent-skills","ai-coding-agents","ai-tools","chain-of-thought","chatgpt","claude-code","claude-skill","cursor","gemini-cli","github-copilot","llm-prompts","openai-codex","prompt-engineering","prompt-frameworks","prompt-improvement","prompt-optimization","roo-code","vscode","windsurf"],"license":"mit","html_url":"https://github.com/ckelsoe/prompt-architect","pushed_at":"2026-03-30T21:30:49Z","description":"Claude Code skill that transforms vague prompts into structured, expert-level prompts using 7 research-backed   frameworks (CO-STAR, RISEN, RISE, TIDD-EC, RTF, CoT, CoD)","skill_md_sha":"f77de814bd8ee0d957570e870bbb0c3c14f1e6c9","skill_md_path":"skills/prompt-architect/SKILL.md","default_branch":"main","skill_tree_url":"https://github.com/ckelsoe/prompt-architect/tree/main/skills/prompt-architect"},"layout":"multi","source":"github","category":"prompt-architect","frontmatter":{"name":"prompt-architect","license":"MIT","description":"Analyzes and improves prompts using 27 research-backed frameworks across 7 intent categories. Use when a user wants to improve, rewrite, structure, or engineer a prompt — including requests like \"help me write a better prompt\", \"improve this prompt\", \"what framework should I use\", \"make this prompt more effective\", or any prompt engineering task. Recommends the right framework based on intent (create, transform, reason, critique, recover, clarify, agentic), asks targeted questions, and delivers a structured, high-quality result.","compatibility":"Requires no external dependencies. Works with any Agent Skills compatible tool."},"skills_sh_url":"https://skills.sh/ckelsoe/prompt-architect/prompt-architect"},"updatedAt":"2026-05-02T12:54:45.789Z"}}