{"id":"6e06ba36-c4d1-4174-af28-9be8b73c2b27","shortId":"e8AGaF","kind":"skill","title":"prd-review-challenger","tagline":"Critical review of PRDs, feature specs, and product decision documents — acts as a devil's advocate to surface weak assumptions, missing requirements, implementation risks, and logical gaps. Use when a PM or team lead wants to stress-test a document before team review, validate a","description":"# Skill: prd-review-challenger\n\nActs as a devil's advocate for PRDs, feature specs, and product decisions. Surfaces\nweak spots, hidden assumptions, logical gaps, and overlooked risks — before the\ndocument goes to the team or into development.\n\n---\n\n## Triggers\n\n**Russian:** «проверь PRD», «разбери спецификацию», «найди дыры в PRD», «покритикуй спецификацию», «ревью PRD», «challenger для PRD», «что упущено в спецификации».\n**English:** \"review my PRD\", \"challenge my spec\",\n\"find holes in my PRD\", \"critique my feature spec\", \"stress-test my PRD\",\n\"what's missing in my spec\"\n\n---\n\n## Language Detection\n\nDetect the language of the input document — respond in that language.\nIf a different language is explicitly requested, use it.\n\n---\n\n## Input\n\n**Required:** text of a PRD, feature spec, or product decision description\n(pasted directly into the conversation or provided as a file).\n\n**Optional:** context that sharpens the critique:\n- product target audience\n- project stage (0→1, scaling, redesign)\n- strategic constraints (timeline, tech stack, budget)\n- the angle that matters most right now (UX, technical complexity, business risks)\n\n## Output\n\nA structured review report with five sections:\n\n1. Weak spots and hidden assumptions\n2. Open questions — what the PRD doesn't answer\n3. Implementation and UX risks\n4. Alternative approaches to consider\n5. Completeness checklist with rating\n\n---\n\n## Instructions\n\n### Step 1 — Accept the Document and Assess Its Type\n\nReceive the document text. Identify the document type:\n\n| Document type | Action |\n|--------------|--------|\n| Full PRD / feature spec | Proceed to Step 2 |\n| Too short (< 3 paragraphs) | State: \"The document is too short for a full review — I'll give feedback on what's available and list what's structurally missing\" |\n| Technical document (not a PRD) | Apply the same critical approach; open with: \"This is a technical document, not a PRD — I'll review it as a specification\" |\n| Request to write a PRD | Decline: this skill critiques, it does not create. Suggest `product-management:write-spec` |\n\n### Step 2 — Completeness Checklist\n\nAssess whether each component is present. This provides the structural foundation\nfor the critique in Steps 3–4.\n\nMark each item: present (✓), absent (✗), partial (△), not applicable (—).\n\n**For PRDs and feature specs:**\n\n- Problem and its scope (problem statement + evidence)\n- Target audience / personas\n- User stories or usage scenarios\n- Acceptance criteria (measurable definition of done)\n- Edge cases and non-standard scenarios\n- Success metrics (how we'll know the feature works)\n- Out of scope (what we explicitly are not doing)\n- Technical constraints and dependencies\n- Rollback / plan B on failure\n- Impact on other parts of the product\n\n**For technical documents (design doc, architecture doc, migration plan) —\nreplace non-applicable items:**\n\n- Problem and its scope (applicable)\n- Owner and reviewers (instead of \"Target audience\")\n- Failure scenarios and boundary states (instead of \"User stories\")\n- Success criteria / test plan (instead of \"Acceptance criteria\")\n- Infrastructure edge cases (applicable)\n- Performance metrics and observability (instead of \"Feature metrics\")\n- Out of scope (applicable)\n- Dependencies and preconditions (applicable)\n- Rollback plan (applicable)\n- Impact on adjacent systems (instead of \"Product impact\")\n\nMark \"—\" only when an item is objectively inapplicable to the document type.\nA missing item is a finding for the critique — not just a formality.\n\n### Step 3 — Four-Axis Critique\n\nProduce the critique as four sections. Each section must contain specific\nobservations with references to the document (direct quote or section reference).\nDo not give generic feedback (\"needs better description\") — give specifics\n(\"section X states Y, but does not explain Z\").\n\n**Section 1: Weak spots and hidden assumptions**\nFind statements treated as facts that have not been validated or justified.\nFormat: \"The document assumes [X] — but this has not been proven / verified /\nconflicts with [Y].\"\n\n**Section 2: Open questions**\nList questions the document doesn't answer but engineers or designers will\ndefinitely ask. Format: \"What happens if [scenario]?\",\n\"How does the system behave when [condition]?\"\n\n**Section 3: Implementation and UX risks**\nIdentify two types of risks:\n- Technical: non-obvious complexity, dependencies, performance, security\n- UX: non-obvious behavior for the user, friction, interface edge cases\n\nFor each risk: description + potential consequence if left unaddressed.\n\n**Section 4: Alternative approaches**\nPropose 2–3 alternatives to the chosen approach — not to replace it, but to\nconfirm the decision is deliberate. Format: \"Alternative: [description] —\ntrade-off: [upside] vs [downside].\"\n\n### Step 4 — Final Output\n\nClose with two blocks:\n\n**Completeness assessment** — the checklist from Step 2 as a table:\ncomponent, status (✓/✗/△/—), brief comment on missing items.\n\n**Top 3 priorities** — the three most critical gaps to close before development\nstarts. Format: numbered list, one sentence per item.\n\n---\n\n## Constraints\n\n- Does not write or rewrite PRDs — critique only\n- Has no domain knowledge about the specific product without provided context\n- Does not replace review with actual stakeholders, engineers, and designers\n- Does not provide legal, financial, or technical expert assessments without\n  relevant context\n- If the PRD is confidential (NDA, internal), the user takes responsibility\n  for sharing the content","tags":["prd","review","challenger","claude","skills","kit","kirkruglov","agent-skills","agentic-skills","ai-agents","ai-skills","awesome-list"],"capabilities":["skill","source-kirkruglov","skill-prd-review-challenger","topic-agent-skills","topic-agentic-skills","topic-ai-agents","topic-ai-skills","topic-awesome-list","topic-claude","topic-claude-ai","topic-claude-ai-skills","topic-claude-code","topic-claude-cowork","topic-claude-memory","topic-claude-skills"],"categories":["claude-skills-kit"],"synonyms":[],"warnings":[],"endpointUrl":"https://skills.sh/KirKruglov/claude-skills-kit/prd-review-challenger","protocol":"skill","transport":"skills-sh","auth":{"type":"none","details":{"cli":"npx skills add KirKruglov/claude-skills-kit","source_repo":"https://github.com/KirKruglov/claude-skills-kit","install_from":"skills.sh"}},"qualityScore":"0.453","qualityRationale":"deterministic score 0.45 from registry signals: · indexed on github topic:agent-skills · 7 github stars · SKILL.md body (5,516 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":"2026-05-18T19:13:38.154Z","embedding":null,"createdAt":"2026-05-18T13:21:15.759Z","updatedAt":"2026-05-18T19:13:38.154Z","lastSeenAt":"2026-05-18T19:13:38.154Z","tsv":"'0':192 '1':193,222,254,600 '2':228,280,357,634,708,749 '3':237,283,376,553,664,709,761 '4':242,377,704,736 '5':247 'absent':382 'accept':255,406,494 'act':15,56 'action':272 'actual':805 'adjac':521 'advoc':20,61 'altern':243,705,710,727 'angl':203 'answer':236,643 'appli':314 'applic':385,465,471,499,511,515,518 'approach':244,318,706,714 'architectur':458 'ask':650 'assess':259,360,744,818 'assum':621 'assumpt':24,73,227,605 'audienc':189,399,478 'avail':302 'axi':556 'b':443 'behav':660 'behavior':686 'better':586 'block':742 'boundari':482 'brief':755 'budget':201 'busi':212 'case':413,498,693 'challeng':4,55,103,114 'checklist':249,359,746 'chosen':713 'close':739,769 'comment':756 'complet':248,358,743 'complex':211,678 'compon':363,753 'condit':662 'confidenti':826 'confirm':721 'conflict':630 'consequ':699 'consid':246 'constraint':197,438,780 'contain':567 'content':836 'context':182,799,821 'convers':175 'creat':348 'criteria':407,489,495 'critic':5,317,766 'critiqu':122,186,344,373,547,557,560,787 'decis':13,68,169,723 'declin':341 'definit':409,649 'deliber':725 'depend':440,512,679 'descript':170,587,697,728 'design':456,647,809 'detect':138,139 'develop':88,771 'devil':18,59 'differ':152 'direct':172,575 'doc':457,459 'document':14,45,81,145,257,264,268,270,287,310,325,455,537,574,620,640 'doesn':234,641 'domain':791 'done':411 'downsid':734 'edg':412,497,692 'engin':645,807 'english':110 'evid':397 'expert':817 'explain':597 'explicit':155,433 'fact':610 'failur':445,479 'featur':9,64,124,165,275,389,426,506 'feedback':298,584 'file':180 'final':737 'financi':814 'find':117,544,606 'five':220 'formal':551 'format':618,651,726,773 'foundat':370 'four':555,562 'four-axi':554 'friction':690 'full':273,293 'gap':31,75,767 'generic':583 'give':297,582,588 'goe':82 'happen':653 'hidden':72,226,604 'hole':118 'identifi':266,669 'impact':446,519,526 'implement':27,238,665 'inapplic':534 'infrastructur':496 'input':144,159 'instead':475,484,492,504,523 'instruct':252 'interfac':691 'intern':828 'item':380,466,531,541,759,779 'justifi':617 'know':424 'knowledg':792 'languag':137,141,149,153 'lead':38 'left':701 'legal':813 'list':304,637,775 'll':296,330,423 'logic':30,74 'manag':352 'mark':378,527 'matter':205 'measur':408 'metric':420,501,507 'migrat':460 'miss':25,133,308,540,758 'must':566 'nda':827 'need':585 'non':416,464,676,684 'non-applic':463 'non-obvi':675,683 'non-standard':415 'number':774 'object':533 'observ':503,569 'obvious':677,685 'one':776 'open':229,319,635 'option':181 'output':214,738 'overlook':77 'owner':472 'paragraph':284 'part':449 'partial':383 'past':171 'per':778 'perform':500,680 'persona':400 'plan':442,461,491,517 'pm':35 'potenti':698 'prd':2,53,92,98,102,105,113,121,130,164,233,274,313,328,340,824 'prd-review-challeng':1,52 'prds':8,63,387,786 'precondit':514 'present':365,381 'prioriti':762 'problem':391,395,467 'proceed':277 'produc':558 'product':12,67,168,187,351,452,525,796 'product-manag':350 'project':190 'propos':707 'proven':628 'provid':177,367,798,812 'question':230,636,638 'quot':576 'rate':251 'receiv':262 'redesign':195 'refer':571,579 'relev':820 'replac':462,717,802 'report':218 'request':156,336 'requir':26,160 'respond':146 'respons':832 'review':3,6,48,54,111,217,294,331,474,803 'rewrit':785 'right':207 'risk':28,78,213,241,668,673,696 'rollback':441,516 'russian':90 'scale':194 'scenario':405,418,480,655 'scope':394,430,470,510 'section':221,563,565,578,590,599,633,663,703 'secur':681 'sentenc':777 'share':834 'sharpen':184 'short':282,290 'skill':51,343 'skill-prd-review-challenger' 'source-kirkruglov' 'spec':10,65,116,125,136,166,276,355,390 'specif':335,568,589,795 'spot':71,224,602 'stack':200 'stage':191 'stakehold':806 'standard':417 'start':772 'state':285,483,592 'statement':396,607 'status':754 'step':253,279,356,375,552,735,748 'stori':402,487 'strateg':196 'stress':42,127 'stress-test':41,126 'structur':216,307,369 'success':419,488 'suggest':349 'surfac':22,69 'system':522,659 'tabl':752 'take':831 'target':188,398,477 'team':37,47,85 'tech':199 'technic':210,309,324,437,454,674,816 'test':43,128,490 'text':161,265 'three':764 'timelin':198 'top':760 'topic-agent-skills' 'topic-agentic-skills' 'topic-ai-agents' 'topic-ai-skills' 'topic-awesome-list' 'topic-claude' 'topic-claude-ai' 'topic-claude-ai-skills' 'topic-claude-code' 'topic-claude-cowork' 'topic-claude-memory' 'topic-claude-skills' 'trade':730 'trade-off':729 'treat':608 'trigger':89 'two':670,741 'type':261,269,271,538,671 'unaddress':702 'upsid':732 'usag':404 'use':32,157 'user':401,486,689,830 'ux':209,240,667,682 'valid':49,615 'verifi':629 'vs':733 'want':39 'weak':23,70,223,601 'whether':361 'without':797,819 'work':427 'write':338,354,783 'write-spec':353 'x':591,622 'y':593,632 'z':598 'в':97,108 'для':104 'дыры':96 'найди':95 'покритикуй':99 'проверь':91 'разбери':93 'ревью':101 'спецификации':109 'спецификацию':94,100 'упущено':107 'что':106","prices":[{"id":"950e8b34-caa4-46d9-b386-8cfbe3147410","listingId":"6e06ba36-c4d1-4174-af28-9be8b73c2b27","amountUsd":"0","unit":"free","nativeCurrency":null,"nativeAmount":null,"chain":null,"payTo":null,"paymentMethod":"skill-free","isPrimary":true,"details":{"org":"KirKruglov","category":"claude-skills-kit","install_from":"skills.sh"},"createdAt":"2026-05-18T13:21:15.759Z"}],"sources":[{"listingId":"6e06ba36-c4d1-4174-af28-9be8b73c2b27","source":"github","sourceId":"KirKruglov/claude-skills-kit/prd-review-challenger","sourceUrl":"https://github.com/KirKruglov/claude-skills-kit/tree/main/skills/prd-review-challenger","isPrimary":false,"firstSeenAt":"2026-05-18T13:21:15.759Z","lastSeenAt":"2026-05-18T19:13:38.154Z"}],"details":{"listingId":"6e06ba36-c4d1-4174-af28-9be8b73c2b27","quickStartSnippet":null,"exampleRequest":null,"exampleResponse":null,"schema":null,"openapiUrl":null,"agentsTxtUrl":null,"citations":[],"useCases":[],"bestFor":[],"notFor":[],"kindDetails":{"org":"KirKruglov","slug":"prd-review-challenger","github":{"repo":"KirKruglov/claude-skills-kit","stars":7,"topics":["agent-skills","agentic-skills","ai","ai-agents","ai-skills","awesome-list","claude","claude-ai","claude-ai-skills","claude-code","claude-cowork","claude-memory","claude-skills","memory-management","productivity","productivity-tools","project-management"],"license":"mit","html_url":"https://github.com/KirKruglov/claude-skills-kit","pushed_at":"2026-05-18T04:27:46Z","description":"40+ curated agent skills for Claude Cowork and Claude.ai — ready-to-use tools for non-technical users: project management, productivity, and AI workflow automation","skill_md_sha":"070b21ff70fe8dd04cd76788d03d7e2a6b460de0","skill_md_path":"skills/prd-review-challenger/SKILL.md","default_branch":"main","skill_tree_url":"https://github.com/KirKruglov/claude-skills-kit/tree/main/skills/prd-review-challenger"},"layout":"multi","source":"github","category":"claude-skills-kit","frontmatter":{"name":"prd-review-challenger","description":"Critical review of PRDs, feature specs, and product decision documents — acts as a devil's advocate to surface weak assumptions, missing requirements, implementation risks, and logical gaps. Use when a PM or team lead wants to stress-test a document before team review, validate a spec before development starts, or identify what's missing in a product decision. Triggers RU: «проверь PRD», «разбери спецификацию», «найди дыры в PRD». Triggers EN: 'review my PRD', 'challenge my spec', 'find holes in my PRD', 'critique my feature spec'. Do not use for writing or rewriting PRDs."},"skills_sh_url":"https://skills.sh/KirKruglov/claude-skills-kit/prd-review-challenger"},"updatedAt":"2026-05-18T19:13:38.154Z"}}