{"id":"a34b8d27-cbf3-48c3-b6a7-16934cd8fabf","shortId":"LEaJB5","kind":"skill","title":"claude","tagline":"Claude Code CLI wrapper for getting an independent second opinion. Three modes:\nReview (diff review via claude -p), Challenge (adversarial failure-mode review),\nConsult (ask Claude about the repo with read-only file tools). Use when asked for\n\"claude review\", \"claude challenge\", ","description":"## Preamble\n\n```bash\neval \"$(~/.vibestack/bin/vibe-slug 2>/dev/null)\" 2>/dev/null || SLUG=\"unknown\"\n_LEARN_FILE=\"${VIBESTACK_HOME:-$HOME/.vibestack}/projects/${SLUG:-unknown}/learnings.jsonl\"\nif [ -f \"$_LEARN_FILE\" ]; then\n  _LEARN_COUNT=$(wc -l < \"$_LEARN_FILE\" 2>/dev/null | tr -d ' ')\n  echo \"LEARNINGS: $_LEARN_COUNT entries loaded\"\n  if [ \"$_LEARN_COUNT\" -gt 5 ] 2>/dev/null; then\n    ~/.vibestack/bin/vibe-learnings-search --limit 5 2>/dev/null || true\n  fi\nelse\n  echo \"LEARNINGS: none yet\"\nfi\n```\n\n## Base Branch\n\n```bash\nBASE_BRANCH=$(git symbolic-ref refs/remotes/origin/HEAD 2>/dev/null | sed 's@^refs/remotes/origin/@@' || echo \"main\")\n```\n\n# /claude - Claude Outside Voice\n\nThis skill wraps `claude -p` to get an independent Claude Code second opinion\nwithout allowing nested Claude to modify files.\n\nThe generated external invocation name is `vibe-claude`.\n\n---\n\n## Step 0: Check Claude CLI\n\n```bash\nCLAUDE_BIN=$(command -v claude 2>/dev/null || echo \"\")\n[ -z \"$CLAUDE_BIN\" ] && echo \"NOT_FOUND\" || echo \"FOUND: $CLAUDE_BIN\"\n```\n\nIf `NOT_FOUND`, stop and tell the user:\n\"Claude CLI not found. Install Claude Code, then re-run this skill.\"\n\nCheck auth:\n\n```bash\nif [ -f \"$HOME/.claude/.credentials.json\" ] || [ -n \"${ANTHROPIC_API_KEY:-}\" ]; then\n  echo \"AUTH_FOUND\"\nelse\n  echo \"AUTH_MISSING\"\nfi\n```\n\nIf `AUTH_MISSING`, stop and tell the user:\n\"No Claude authentication found. Run `claude` interactively to log in, or export `ANTHROPIC_API_KEY`, then re-run this skill.\"\n\n---\n\n## Safety Boundary\n\nNested Claude must stay focused on the user's repository and must not run\nvibestack skills from inside this skill.\n\nAll `claude -p` calls MUST include:\n\n- `--disable-slash-commands`\n- Review/challenge: `--tools \"\"`\n- Consult: `--allowedTools Read,Grep,Glob --disallowedTools Bash,Edit,Write`\n\nNever pass `Bash`, `Edit`, or `Write` to nested Claude in this skill.\n\nAll prompts MUST be written to a temp file and fed through stdin. Never interpolate\nuser text directly into the shell command.\n\n---\n\n## Step 1: Detect Mode\n\nParse the user's input:\n\n1. `/claude review` or `/claude review <instructions>` - **Review mode** (Step 2A)\n2. `/claude challenge` or `/claude challenge <focus>` - **Challenge mode** (Step 2B)\n3. `/claude` with no arguments, or `/claude <anything else>` - **Consult mode** (Step 2C)\n\nIf no mode is obvious and a diff exists, ask whether to review, challenge, or consult.\n\n---\n\n## Shared Helpers\n\nUse these shell snippets in every mode.\n\nCreate temp files:\n\n```bash\nPROMPT_FILE=$(mktemp /tmp/vibe-claude-prompt-XXXXXX)\nRESP_FILE=$(mktemp /tmp/vibe-claude-response-XXXXXX.json)\nERR_FILE=$(mktemp /tmp/vibe-claude-error-XXXXXX.txt)\n```\n\nCleanup at the end of every mode:\n\n```bash\nrm -f \"$PROMPT_FILE\" \"$RESP_FILE\" \"$ERR_FILE\"\n```\n\nParse JSON output:\n\n```bash\npython3 - \"$RESP_FILE\" <<'PY'\nimport json, sys\npath = sys.argv[1]\ntry:\n    obj = json.load(open(path))\nexcept Exception as exc:\n    print(f\"CLAUDE_JSON_PARSE_ERROR: {exc}\")\n    sys.exit(0)\n\nif obj.get(\"is_error\"):\n    print(\"CLAUDE_ERROR: true\")\n\nresult = obj.get(\"result\") or obj.get(\"response\") or \"\"\nif result:\n    print(result)\n\nusage = obj.get(\"usage\") or {}\ninput_tokens = usage.get(\"input_tokens\", 0) or 0\noutput_tokens = usage.get(\"output_tokens\", 0) or 0\ncache_read = usage.get(\"cache_read_input_tokens\", 0) or 0\nmodel = obj.get(\"model\") or \"unknown\"\nsession_id = obj.get(\"session_id\") or \"\"\n\nprint(f\"\\nTokens: input={input_tokens} output={output_tokens} cache_read={cache_read} | Model: {model}\")\nif session_id:\n    print(f\"SESSION_ID:{session_id}\")\nPY\n```\n\nIf stderr contains `auth`, `login`, or `unauthorized`, tell the user:\n\"Claude authentication failed. Run `claude` interactively to authenticate or export `ANTHROPIC_API_KEY`.\"\n\n---\n\n## Step 2A: Review Mode\n\nReview the current branch diff with nested Claude in tool-less mode.\n\n1. Fetch base and capture diff:\n\n```bash\n_REPO_ROOT=$(git rev-parse --show-toplevel) || { echo \"ERROR: not in a git repo\" >&2; exit 1; }\ncd \"$_REPO_ROOT\"\nDIFF_FILE=$(mktemp /tmp/vibe-claude-diff-XXXXXX.patch)\ngit fetch origin \"$BASE_BRANCH\" --quiet 2>/dev/null || true\ngit diff \"origin/$BASE_BRANCH\" > \"$DIFF_FILE\" 2>/dev/null || git diff \"$BASE_BRANCH\" > \"$DIFF_FILE\"\n```\n\nIf the diff file is empty, stop and say:\n\"Nothing to review - no changes against the base branch.\"\n\n2. Write the prompt file:\n\n```bash\ncat > \"$PROMPT_FILE\" <<'EOF'\nYou are a brutally honest Claude Code reviewer. Review this git diff for bugs,\nproduction failure modes, security issues, missing tests, and maintainability\nproblems. Be direct. No compliments. Reference files and changed code where possible.\n\nAdditional user instructions, if any:\n<custom review instructions>\n\nDIFF:\nEOF\ncat \"$DIFF_FILE\" >> \"$PROMPT_FILE\"\n```\n\n3. Run Claude:\n\n```bash\ncat \"$PROMPT_FILE\" | claude -p --output-format json --disable-slash-commands --tools \"\" > \"$RESP_FILE\" 2>\"$ERR_FILE\"\n```\n\n4. Present the parsed output:\n\n```\nCLAUDE SAYS (code review):\n============================================================\n<parsed result from RESP_FILE>\n============================================================\n```\n\n5. Cleanup:\n\n```bash\nrm -f \"$DIFF_FILE\" \"$PROMPT_FILE\" \"$RESP_FILE\" \"$ERR_FILE\"\n```\n\n---\n\n## Step 2B: Challenge Mode\n\nRun an adversarial failure-mode review with nested Claude in tool-less mode.\n\n1. Capture the diff using the same diff commands from Review mode.\n\n2. Write the prompt:\n\n```bash\ncat > \"$PROMPT_FILE\" <<'EOF'\nYou are an adversarial Claude Code reviewer. Try to break this change before users do.\nFind edge cases, race conditions, security holes, resource leaks, silent data\ncorruption, bad error handling, and operational failure modes. Be thorough. No\ncompliments. If the user provided a focus area, prioritize it.\n\nFocus area, if any:\n<focus>\n\nDIFF:\nEOF\ncat \"$DIFF_FILE\" >> \"$PROMPT_FILE\"\n```\n\n3. Run Claude:\n\n```bash\ncat \"$PROMPT_FILE\" | claude -p --output-format json --disable-slash-commands --tools \"\" > \"$RESP_FILE\" 2>\"$ERR_FILE\"\n```\n\n4. Present the parsed output:\n\n```\nCLAUDE SAYS (adversarial challenge):\n============================================================\n<parsed result from RESP_FILE>\n============================================================\n```\n\n5. Cleanup:\n\n```bash\nrm -f \"$DIFF_FILE\" \"$PROMPT_FILE\" \"$RESP_FILE\" \"$ERR_FILE\"\n```\n\n---\n\n## Step 2C: Consult Mode\n\nAsk Claude about the repository. Consult mode may inspect files, but only with\nread-only tools.\n\n1. Check for an existing Claude session:\n\n```bash\ncat .context/claude-session-id 2>/dev/null || echo \"NO_SESSION\"\n```\n\nIf a session exists, ask the user whether to continue it or start fresh.\n\n2. Write the prompt:\n\n```bash\ncat > \"$PROMPT_FILE\" <<'EOF'\nYou are Claude Code acting as an independent outside voice for this repository.\nAnswer the user's question directly. You may inspect repository files with Read,\nGrep, and Glob only. Do not use Bash. Do not edit or write files. Do not invoke\nslash commands or vibestack skills.\n\nUSER QUESTION:\n<user prompt>\nEOF\n```\n\n3. Run Claude.\n\nFor a new session:\n\n```bash\ncat \"$PROMPT_FILE\" | claude -p --output-format json --disable-slash-commands --allowedTools Read,Grep,Glob --disallowedTools Bash,Edit,Write > \"$RESP_FILE\" 2>\"$ERR_FILE\"\n```\n\nFor a resumed session:\n\n```bash\ncat \"$PROMPT_FILE\" | claude -p --resume \"<session-id>\" --output-format json --disable-slash-commands --allowedTools Read,Grep,Glob --disallowedTools Bash,Edit,Write > \"$RESP_FILE\" 2>\"$ERR_FILE\"\n```\n\n4. Parse and save the session id:\n\n```bash\nSESSION_ID=$(python3 - \"$RESP_FILE\" <<'PY'\nimport json, sys\ntry:\n    obj = json.load(open(sys.argv[1]))\n    print(obj.get(\"session_id\") or \"\")\nexcept Exception:\n    print(\"\")\nPY\n)\nif [ -n \"$SESSION_ID\" ]; then\n  mkdir -p .context\n  printf \"%s\\n\" \"$SESSION_ID\" > .context/claude-session-id\nfi\n```\n\n5. Present the parsed output:\n\n```\nCLAUDE SAYS (consult):\n============================================================\n<parsed result from RESP_FILE>\n============================================================\nSession saved - run /claude again to continue this conversation.\n```\n\n6. Cleanup:\n\n```bash\nrm -f \"$PROMPT_FILE\" \"$RESP_FILE\" \"$ERR_FILE\"\n```\n\n---\n\n## Error Handling\n\n- **Binary not found:** Stop with install instructions.\n- **Auth missing:** Stop with login/API key instructions.\n- **Auth failure from stderr:** Surface the stderr line and ask the user to re-authenticate.\n- **JSON parse failure:** Show raw stdout from `$RESP_FILE` and stderr from `$ERR_FILE`.\n- **Empty response:** Tell the user \"Claude returned no response. Check stderr for errors.\"\n- **Resume failure:** Delete `.context/claude-session-id` and retry with a fresh session.\n\n---\n\n## Important Rules\n\n- Nested Claude is read-only in consult mode and tool-less in review/challenge.\n- Always include `--disable-slash-commands`.\n- Never pass nested Claude `Bash`, `Edit`, or `Write`.\n- Never interpolate user text into a shell command.\n- Present Claude's response faithfully, then add any host-agent synthesis after it.","tags":["claude","vibestack","timurgaleev","agent-skills","ai-agents","claude-code","cursor-ide","developer-tools","kiro","mcp","prompt-engineering","slash-commands"],"capabilities":["skill","source-timurgaleev","skill-claude","topic-agent-skills","topic-ai-agents","topic-claude-code","topic-cursor-ide","topic-developer-tools","topic-kiro","topic-mcp","topic-prompt-engineering","topic-slash-commands"],"categories":["vibestack"],"synonyms":[],"warnings":[],"endpointUrl":"https://skills.sh/timurgaleev/vibestack/claude","protocol":"skill","transport":"skills-sh","auth":{"type":"none","details":{"cli":"npx skills add timurgaleev/vibestack","source_repo":"https://github.com/timurgaleev/vibestack","install_from":"skills.sh"}},"qualityScore":"0.457","qualityRationale":"deterministic score 0.46 from registry signals: · indexed on github topic:agent-skills · 15 github stars · SKILL.md body (8,952 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":"2026-05-18T19:06:19.889Z","embedding":null,"createdAt":"2026-05-18T19:06:19.889Z","updatedAt":"2026-05-18T19:06:19.889Z","lastSeenAt":"2026-05-18T19:06:19.889Z","tsv":"'/.vibestack/bin/vibe-learnings-search':94 '/.vibestack/bin/vibe-slug':49 '/claude':124,337,340,347,350,357,362,1135 '/dev/null':51,53,77,92,98,118,169,621,631,933 '/learnings.jsonl':64 '/projects':61 '/tmp/vibe-claude-diff-xxxxxx.patch':613 '/tmp/vibe-claude-error-xxxxxx.txt':407 '/tmp/vibe-claude-prompt-xxxxxx':399 '/tmp/vibe-claude-response-xxxxxx.json':403 '0':158,455,484,486,492,494,502,504 '1':328,336,437,581,606,777,922,1099 '2':50,52,76,91,97,117,168,346,604,620,630,656,733,789,876,932,951,1042,1074 '2a':345,565 '2b':355,759 '2c':366,902 '3':356,713,856,1011 '4':736,879,1077 '5':90,96,745,888,1124 '6':1141 'act':964 'add':1266 'addit':701 'adversari':21,764,801,886 'agent':1270 'allow':142 'allowedtool':285,1032,1064 'alway':1238 'answer':973 'anthrop':209,241,561 'api':210,242,562 'area':842,846 'argument':360 'ask':27,40,376,905,941,1177 'auth':203,214,218,222,544,1161,1168 'authent':231,552,558,1183 'bad':825 'base':107,110,583,617,626,634,654 'bash':47,109,162,204,290,295,395,415,427,587,661,716,747,793,859,890,929,955,993,1018,1037,1049,1069,1084,1143,1248 'bin':164,173,180 'binari':1154 'boundari':251 'branch':108,111,571,618,627,635,655 'break':807 'brutal':669 'bug':679 'cach':495,498,525,527 'call':275 'captur':585,778 'case':815 'cat':662,708,717,794,851,860,930,956,1019,1050 'cd':607 'challeng':20,45,348,351,352,380,760,887 'chang':651,697,809 'check':159,202,923,1207 'claud':1,2,18,28,42,44,125,131,137,144,156,160,163,167,172,179,189,194,230,234,253,273,301,449,461,551,555,575,671,715,720,741,771,802,858,863,884,906,927,962,1013,1022,1053,1129,1203,1224,1247,1261 'cleanup':408,746,889,1142 'cli':4,161,190 'code':3,138,195,672,698,743,803,963 'command':165,281,326,729,785,872,1004,1031,1063,1243,1259 'compliment':693,835 'condit':817 'consult':26,284,363,382,903,910,1131,1230 'contain':543 'context':1116 'context/claude-session-id':931,1122,1214 'continu':946,1138 'convers':1140 'corrupt':824 'count':71,83,88 'creat':392 'current':570 'd':79 'data':823 'delet':1213 'detect':329 'diff':15,374,572,586,610,624,628,633,636,640,677,706,709,750,780,784,849,852,893 'direct':322,691,978 'disabl':279,727,870,1029,1061,1241 'disable-slash-command':278,726,869,1028,1060,1240 'disallowedtool':289,1036,1068 'echo':80,102,122,170,174,177,213,217,597,934 'edg':814 'edit':291,296,996,1038,1070,1249 'els':101,216 'empti':643,1198 'end':411 'entri':84 'eof':665,707,797,850,959,1010 'err':404,422,734,756,877,899,1043,1075,1150,1196 'error':452,459,462,598,826,1152,1210 'eval':48 'everi':390,413 'exc':446,453 'except':443,444,1105,1106 'exist':375,926,940 'exit':605 'export':240,560 'extern':150 'f':66,206,417,448,517,535,749,892,1145 'fail':553 'failur':23,681,766,830,1169,1186,1212 'failure-mod':22,765 'faith':1264 'fed':315 'fetch':582,615 'fi':100,106,220,1123 'file':36,57,68,75,147,313,394,397,401,405,419,421,423,430,611,629,637,641,660,664,695,710,712,719,732,735,751,753,755,757,796,853,855,862,875,878,894,896,898,900,914,958,983,999,1021,1041,1044,1052,1073,1076,1089,1147,1149,1151,1192,1197 'find':813 'focus':256,841,845 'format':724,867,1026,1058 'found':176,178,183,192,215,232,1156 'fresh':950,1219 'generat':149 'get':7,134 'git':112,590,602,614,623,632,676 'glob':288,988,1035,1067 'grep':287,986,1034,1066 'gt':89 'handl':827,1153 'helper':384 'hole':819 'home':59 'home/.claude/.credentials.json':207 'home/.vibestack':60 'honest':670 'host':1269 'host-ag':1268 'id':511,514,533,537,539,1083,1086,1103,1112,1121 'import':432,1091,1221 'includ':277,1239 'independ':9,136,967 'input':335,479,482,500,519,520 'insid':269 'inspect':913,981 'instal':193,1159 'instruct':703,1160,1167 'interact':235,556 'interpol':319,1253 'invoc':151 'invok':1002 'issu':684 'json':425,433,450,725,868,1027,1059,1092,1184 'json.load':440,1096 'key':211,243,563,1166 'l':73 'leak':821 'learn':56,67,70,74,81,82,87,103 'less':579,775,1235 'limit':95 'line':1175 'load':85 'log':237 'login':545 'login/api':1165 'main':123 'maintain':688 'may':912,980 'miss':219,223,685,1162 'mkdir':1114 'mktemp':398,402,406,612 'mode':13,24,330,343,353,364,369,391,414,567,580,682,761,767,776,788,831,904,911,1231 'model':505,507,529,530 'modifi':146 'must':254,263,276,307 'n':208,1110,1119 'name':152 'nest':143,252,300,574,770,1223,1246 'never':293,318,1244,1252 'new':1016 'none':104 'noth':647 'ntoken':518 'obj':439,1095 'obj.get':457,465,468,476,506,512,1101 'obvious':371 'open':441,1097 'oper':829 'opinion':11,140 'origin':616,625 'output':426,487,490,522,523,723,740,866,883,1025,1057,1128 'output-format':722,865,1024,1056 'outsid':126,968 'p':19,132,274,721,864,1023,1054,1115 'pars':331,424,451,593,739,882,1078,1127,1185 'pass':294,1245 'path':435,442 'possibl':700 'preambl':46 'present':737,880,1125,1260 'print':447,460,473,516,534,1100,1107 'printf':1117 'priorit':843 'problem':689 'product':680 'prompt':306,396,418,659,663,711,718,752,792,795,854,861,895,954,957,1020,1051,1146 'provid':839 'py':431,540,1090,1108 'python3':428,1087 'question':977,1009 'quiet':619 'race':816 'raw':1188 're':198,246,1182 're-authent':1181 're-run':197,245 'read':34,286,496,499,526,528,919,985,1033,1065,1227 'read-on':33,918,1226 'ref':115 'refer':694 'refs/remotes/origin':121 'refs/remotes/origin/head':116 'repo':31,588,603,608 'repositori':261,909,972,982 'resourc':820 'resp':400,420,429,731,754,874,897,1040,1072,1088,1148,1191 'respons':469,1199,1206,1263 'result':464,466,472,474 'resum':1047,1055,1211 'retri':1216 'return':1204 'rev':592 'rev-pars':591 'review':14,16,25,43,338,341,342,379,566,568,649,673,674,744,768,787,804 'review/challenge':282,1237 'rm':416,748,891,1144 'root':589,609 'rule':1222 'run':199,233,247,265,554,714,762,857,1012,1134 'safeti':250 'save':1080,1133 'say':646,742,885,1130 'second':10,139 'secur':683,818 'sed':119 'session':510,513,532,536,538,928,936,939,1017,1048,1082,1085,1102,1111,1120,1132,1220 'share':383 'shell':325,387,1258 'show':595,1187 'show-toplevel':594 'silent':822 'skill':129,201,249,267,271,304,1007 'skill-claude' 'slash':280,728,871,1003,1030,1062,1242 'slug':54,62 'snippet':388 'source-timurgaleev' 'start':949 'stay':255 'stderr':542,1171,1174,1194,1208 'stdin':317 'stdout':1189 'step':157,327,344,354,365,564,758,901 'stop':184,224,644,1157,1163 'surfac':1172 'symbol':114 'symbolic-ref':113 'synthesi':1271 'sys':434,1093 'sys.argv':436,1098 'sys.exit':454 'tell':186,226,548,1200 'temp':312,393 'test':686 'text':321,1255 'thorough':833 'three':12 'token':480,483,488,491,501,521,524 'tool':37,283,578,730,774,873,921,1234 'tool-less':577,773,1233 'topic-agent-skills' 'topic-ai-agents' 'topic-claude-code' 'topic-cursor-ide' 'topic-developer-tools' 'topic-kiro' 'topic-mcp' 'topic-prompt-engineering' 'topic-slash-commands' 'toplevel':596 'tr':78 'tri':438,805,1094 'true':99,463,622 'unauthor':547 'unknown':55,63,509 'usag':475,477 'usage.get':481,489,497 'use':38,385,781,992 'user':188,228,259,320,333,550,702,811,838,943,975,1008,1179,1202,1254 'v':166 'via':17 'vibe':155 'vibe-claud':154 'vibestack':58,266,1006 'voic':127,969 'wc':72 'whether':377,944 'without':141 'wrap':130 'wrapper':5 'write':292,298,657,790,952,998,1039,1071,1251 'written':309 'yet':105 'z':171","prices":[{"id":"38e22b2d-a325-41a9-b7f9-94520b915aad","listingId":"a34b8d27-cbf3-48c3-b6a7-16934cd8fabf","amountUsd":"0","unit":"free","nativeCurrency":null,"nativeAmount":null,"chain":null,"payTo":null,"paymentMethod":"skill-free","isPrimary":true,"details":{"org":"timurgaleev","category":"vibestack","install_from":"skills.sh"},"createdAt":"2026-05-18T19:06:19.889Z"}],"sources":[{"listingId":"a34b8d27-cbf3-48c3-b6a7-16934cd8fabf","source":"github","sourceId":"timurgaleev/vibestack/claude","sourceUrl":"https://github.com/timurgaleev/vibestack/tree/main/skills/claude","isPrimary":false,"firstSeenAt":"2026-05-18T19:06:19.889Z","lastSeenAt":"2026-05-18T19:06:19.889Z"}],"details":{"listingId":"a34b8d27-cbf3-48c3-b6a7-16934cd8fabf","quickStartSnippet":null,"exampleRequest":null,"exampleResponse":null,"schema":null,"openapiUrl":null,"agentsTxtUrl":null,"citations":[],"useCases":[],"bestFor":[],"notFor":[],"kindDetails":{"org":"timurgaleev","slug":"claude","github":{"repo":"timurgaleev/vibestack","stars":15,"topics":["agent-skills","ai-agents","claude-code","cursor-ide","developer-tools","kiro","mcp","prompt-engineering","slash-commands"],"license":"mit","html_url":"https://github.com/timurgaleev/vibestack","pushed_at":"2026-05-18T18:19:05Z","description":"vibestack is a portable skill pack for AI coding agents. Slash commands like /office-hours, /ship, /investigate, /tdd, /review install once and work across every agent that supports the Agent Skills open standard — Claude Code, Cursor, Kiro, and a growing list of others. ","skill_md_sha":"ef080c0fddc3f6452fc0374b8779fe913ef7e00d","skill_md_path":"skills/claude/SKILL.md","default_branch":"main","skill_tree_url":"https://github.com/timurgaleev/vibestack/tree/main/skills/claude"},"layout":"multi","source":"github","category":"vibestack","frontmatter":{"name":"claude","description":"Claude Code CLI wrapper for getting an independent second opinion. Three modes:\nReview (diff review via claude -p), Challenge (adversarial failure-mode review),\nConsult (ask Claude about the repo with read-only file tools). Use when asked for\n\"claude review\", \"claude challenge\", \"ask claude\", \"second opinion from claude\",\nor \"outside voice\"."},"skills_sh_url":"https://skills.sh/timurgaleev/vibestack/claude"},"updatedAt":"2026-05-18T19:06:19.889Z"}}