{"id":"65c72974-bbbf-4d9a-97d6-dddb8a0e6178","shortId":"fKH53a","kind":"skill","title":"gemini","tagline":"Use when the user asks to run Gemini CLI for code review, plan review, or big context (>200k) processing. Ideal for comprehensive analysis requiring large context windows. Uses Gemini 3 Pro by default for state-of-the-art reasoning and coding.","description":"# Gemini Skill Guide\n\n## When to Use Gemini\n- WHEN ASKED TO BE ACTIVATED\n- **Code Review**: Comprehensive code reviews across multiple files\n- **Plan Review**: Analyzing architectural plans, technical specifications, or project roadmaps\n- **Big Context Processing**: Tasks requiring >200k tokens of context (entire codebases, documentation sets)\n- **Multi-file Analysis**: Understanding relationships and patterns across many files\n\n## ⚠️ Critical: Background/Non-Interactive Mode Warning\n\n**NEVER use `--approval-mode default` in background or non-interactive shells** (like Claude Code tool calls). It will hang indefinitely waiting for approval prompts that cannot be provided.\n\n**For automated/background reviews:**\n- ✅ Use `--approval-mode yolo` for fully automated execution\n- ✅ OR wrap with timeout: `timeout 300 gemini ...`\n- ❌ NEVER use `--approval-mode default` without interactive terminal\n\n**Symptoms of hung Gemini:**\n- Process running 20+ minutes with 0% CPU usage\n- No network activity\n- Process state shows 'S' (sleeping)\n\n**Fix hung process:**\n```bash\n# Check if hung\nps aux | grep gemini | grep -v grep\n\n# Kill if necessary\npkill -9 -f \"gemini.*gemini-3-pro-preview\"\n```\n\n## Running a Task\n\n1. Ask the user (via `AskUserQuestion`) which model to use in a **single prompt**. Available models:\n   - `gemini-3-pro-preview` ⭐ (flagship model, best for coding & complex reasoning, 35% better at software engineering than 2.5 Pro)\n   - `gemini-3-flash` (sub-second latency, distilled from 3 Pro, best for speed-critical tasks)\n   - `gemini-2.5-pro` (legacy option, strong all-around performance)\n   - `gemini-2.5-flash` (legacy option, cost-efficient with thinking capabilities)\n   - `gemini-2.5-flash-lite` (legacy option, fastest processing)\n\n2. Select the approval mode based on the task:\n   - `default`: Prompt for approval (⚠️ ONLY for interactive terminal sessions)\n   - `auto_edit`: Auto-approve edit tools only (for code reviews with suggestions)\n   - `yolo`: Auto-approve all tools (✅ REQUIRED for background/automated tasks)\n\n3. Assemble the command with appropriate options:\n   - `-m, --model <MODEL>` - Model selection\n   - `--approval-mode <default|auto_edit|yolo>` - Control tool approval\n   - `-y, --yolo` - Alternative to `--approval-mode yolo`\n   - `-i, --prompt-interactive \"prompt\"` - Execute prompt and continue interactively\n   - `--include-directories <DIR>` - Additional directories to include in workspace\n   - `-s, --sandbox` - Run in sandbox mode for isolation\n\n4. **For background/automated tasks, ALWAYS use `--approval-mode yolo`** or add timeout wrapper. NEVER use `default` in non-interactive shells.\n\n5. Run the command and capture the output. For background/automated mode:\n   ```bash\n   # Recommended: Use yolo for background tasks\n   gemini -m gemini-3-pro-preview --approval-mode yolo \"Review this codebase for security issues\"\n\n   # Or with timeout (5 min limit)\n   timeout 300 gemini -m gemini-3-pro-preview --approval-mode yolo \"Review this codebase\"\n   ```\n\n6. For interactive sessions with an initial prompt:\n   ```bash\n   gemini -m gemini-3-pro-preview -i \"Review the authentication system\" --approval-mode auto_edit\n   ```\n\n7. **After Gemini completes**, inform the user: \"The Gemini analysis is complete. You can start a new Gemini session for follow-up analysis or continue exploring the findings.\"\n\n### Quick Reference\n\n| Use case | Approval mode | Key flags |\n| --- | --- | --- |\n| Background code review | `yolo` ✅ | `-m gemini-3-pro-preview --approval-mode yolo` |\n| Background analysis | `yolo` ✅ | `-m gemini-3-pro-preview --approval-mode yolo` |\n| Background with timeout | `yolo` ✅ | `timeout 300 gemini -m gemini-3-pro-preview --approval-mode yolo` |\n| Interactive code review | `default` | `-m gemini-3-pro-preview --approval-mode default` (interactive terminal only) |\n| Code review with auto-edits | `auto_edit` | `-m gemini-3-pro-preview --approval-mode auto_edit` |\n| Automated refactoring | `yolo` | `-m gemini-3-pro-preview --approval-mode yolo` |\n| Speed-critical background | `yolo` ✅ | `-m gemini-3-flash --approval-mode yolo` |\n| Cost-optimized background | `yolo` ✅ | `-m gemini-2.5-flash --approval-mode yolo` |\n| Multi-directory analysis | `yolo` (if background) | `--include-directories <DIR1> --include-directories <DIR2>` |\n| Interactive with prompt | `auto_edit` or `default` | `-i \"prompt\" --approval-mode <mode>` |\n\n### Model Selection Guide\n\n| Model | Best for | Context window | Key features |\n| --- | --- | --- | --- |\n| `gemini-3-pro-preview` ⭐ | **Flagship model**: Complex reasoning, coding, agentic tasks | 1M input / 64k output | Vibe coding, 76.2% SWE-bench, $2-4/M input |\n| `gemini-3-flash` | Sub-second latency, speed-critical applications | 1M input / 64k output | Distilled from 3 Pro, TPU-optimized |\n| `gemini-2.5-pro` | Legacy: Strong all-around performance | 1M input / 65k output | Thinking mode, mature stability |\n| `gemini-2.5-flash` | Legacy: Cost-efficient, high-volume tasks | 1M input / 65k output | Best price ($0.15/M), thinking mode |\n| `gemini-2.5-flash-lite` | Legacy: Fastest processing, high throughput | 1M input / 65k output | Maximum speed, minimal latency |\n\n**Gemini 3 Advantages**: 35% higher accuracy in software engineering, state-of-the-art on SWE-bench (76.2%), GPQA Diamond (91.9%), and WebDev Arena (1487 Elo). Knowledge cutoff: January 2025.\n\n**Coming Soon**: `gemini-3-deep-think` for ultra-complex reasoning with enhanced thinking capabilities.\n\n## Common Use Cases\n\n### Code Review (Background/Automated)\n```bash\n# For background execution (Claude Code, CI/CD, etc.)\ngemini -m gemini-3-pro-preview --approval-mode yolo \\\n  \"Perform a comprehensive code review focusing on:\n   1. Security vulnerabilities\n   2. Performance issues\n   3. Code quality and maintainability\n   4. Best practices violations\"\n\n# With timeout safety (5 minutes)\ntimeout 300 gemini -m gemini-3-pro-preview --approval-mode yolo \\\n  \"Perform a comprehensive code review...\"\n```\n\n### Plan Review (Background/Automated)\n```bash\n# For background execution\ngemini -m gemini-3-pro-preview --approval-mode yolo \\\n  \"Review this architectural plan for:\n   1. Scalability concerns\n   2. Missing components\n   3. Integration challenges\n   4. Alternative approaches\"\n```\n\n### Big Context Analysis (Background/Automated)\n```bash\n# For background execution\ngemini -m gemini-3-pro-preview --approval-mode yolo \\\n  \"Analyze the entire codebase to understand:\n   1. Overall architecture\n   2. Key patterns and conventions\n   3. Potential technical debt\n   4. Refactoring opportunities\"\n```\n\n### Interactive Code Review (Terminal Only)\n```bash\n# ONLY use default mode in interactive terminal\ngemini -m gemini-3-pro-preview --approval-mode default \\\n  \"Review the authentication flow for security issues\"\n```\n\n## Following Up\n\n- Gemini CLI sessions are typically one-shot or interactive. Unlike Codex, there's no built-in resume functionality.\n- For follow-up analysis, start a new Gemini session with context from previous findings.\n- When proposing follow-up actions, restate the chosen model and approval mode.\n- Use `AskUserQuestion` after each Gemini command to confirm next steps or gather clarifications.\n\n## Error Handling\n\n- Stop and report failures whenever `gemini --version` or a Gemini command exits non-zero.\n- Request direction before retrying failed commands.\n- Before using high-impact flags (`--approval-mode yolo`, `-y`, `--sandbox`), ask the user for permission using `AskUserQuestion` unless already granted.\n- When output includes warnings or partial results, summarize them and ask how to adjust using `AskUserQuestion`.\n\n## Troubleshooting Hung Gemini Processes\n\n### Detection\n```bash\n# Check for hung processes\nps aux | grep -E \"gemini.*gemini-3\" | grep -v grep\n\n# Look for these symptoms:\n# - Process running 20+ minutes\n# - CPU usage at 0%\n# - Process state 'S' (sleeping)\n# - No network connections\n```\n\n### Diagnosis\n```bash\n# Get detailed process info\nps -o pid,etime,pcpu,stat,command -p <PID>\n\n# Check network activity\nlsof -p <PID> 2>/dev/null | grep -E \"(TCP|ESTABLISHED)\" | wc -l\n# If result is 0, process is hung\n```\n\n### Resolution\n```bash\n# Kill hung Gemini processes\npkill -9 -f \"gemini.*gemini-3-pro-preview\"\n\n# Or kill specific PID\nkill -9 <PID>\n\n# Verify cleanup\nps aux | grep gemini | grep -v grep\n```\n\n### Prevention\n- **ALWAYS use `--approval-mode yolo` for background/automated tasks**\n- Add timeout wrapper for safety: `timeout 300 gemini ...`\n- Never use `--approval-mode default` in non-interactive shells\n- Monitor first run with `ps` to ensure process completes\n\n## Tips for Large Context Processing\n\n1. **Be specific**: Provide clear, structured prompts for what to analyze\n2. **Use include-directories**: Explicitly specify all relevant directories\n3. **Choose the right model**:\n   - Use `gemini-3-pro-preview` for complex reasoning, coding tasks, and maximum analysis quality (recommended default)\n   - Use `gemini-3-flash` for speed-critical tasks requiring sub-second response times\n   - Use `gemini-2.5-flash` for cost-optimized high-volume processing\n4. **Leverage Gemini 3's strengths**: 35% better at software engineering tasks, exceptional at agentic workflows and vibe coding\n5. **Break down complex tasks**: Even with large context, structured analysis is more effective\n6. **Save findings**: Ask Gemini to output structured reports that can be saved for reference\n\n## CLI Version\n\nRequires Gemini CLI v0.16.0 or later for Gemini 3 model support. Check version: `gemini --version`","tags":["gemini","agent","toolkit","softaworks","agent-skills","automation","claude","claude-code","coding-agent","development"],"capabilities":["skill","source-softaworks","skill-gemini","topic-agent-skills","topic-automation","topic-claude","topic-claude-code","topic-coding-agent","topic-development"],"categories":["agent-toolkit"],"synonyms":[],"warnings":[],"endpointUrl":"https://skills.sh/softaworks/agent-toolkit/gemini","protocol":"skill","transport":"skills-sh","auth":{"type":"none","details":{"cli":"npx skills add softaworks/agent-toolkit","source_repo":"https://github.com/softaworks/agent-toolkit","install_from":"skills.sh"}},"qualityScore":"0.700","qualityRationale":"deterministic score 0.70 from registry signals: · indexed on github topic:agent-skills · 1847 github stars · SKILL.md body (9,533 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":"2026-05-18T18:52:57.591Z","embedding":null,"createdAt":"2026-04-18T20:27:52.449Z","updatedAt":"2026-05-18T18:52:57.591Z","lastSeenAt":"2026-05-18T18:52:57.591Z","tsv":"'-2.5':263,273,284,644,734,751,772,1342 '-3':202,226,246,432,457,480,537,550,567,581,602,616,631,686,712,823,853,893,916,952,997,1152,1220,1310,1327 '-4':708 '-9':198,1216,1229 '/dev/null':1195 '/m':709,768 '0':169,1167,1205 '0.15':767 '1':209,868,929,966,1282 '1487':814 '1m':697,722,742,761,781 '2':292,707,871,932,969,1194,1293 '2.5':243 '20':166,1162 '200k':19,79 '2025':819 '3':31,254,333,728,790,874,935,974,1303,1355,1410 '300':149,453,563,889,1255 '35':237,792,1358 '4':389,879,938,978,1352 '5':411,449,886,1371 '6':468,1385 '64k':699,724 '65k':744,763,783 '7':494 '76.2':703,807 '91.9':810 'accuraci':794 'across':61,95 'action':1054 'activ':55,174,1191 'add':400,1249 'addit':375 'adjust':1133 'advantag':791 'agent':695,1366 'all-around':268,738 'alreadi':1118 'altern':356,939 'alway':393,1240 'analysi':24,90,503,517,546,653,943,1038,1321,1381 'analyz':66,960,1292 'applic':721 'approach':940 'appropri':338 'approv':105,126,137,154,295,304,314,326,345,353,359,396,437,462,490,527,542,555,572,586,607,621,634,647,673,858,898,921,957,1002,1060,1105,1243,1260 'approval-mod':104,136,153,344,358,395,436,461,489,541,554,571,585,606,620,633,646,672,857,897,920,956,1001,1104,1242,1259 'architectur':67,926,968 'arena':813 'around':270,740 'art':40,802 'ask':6,52,210,1110,1130,1388 'askuserquest':214,1063,1116,1135 'assembl':334 'authent':487,1007 'auto':310,313,325,348,492,596,598,609,666 'auto-approv':312,324 'auto-edit':595 'autom':142,611 'automated/background':133 'aux':188,1147,1233 'avail':223 'background':109,427,531,545,558,627,640,656,844,911,947 'background/automated':331,391,420,841,908,944,1247 'background/non-interactive':99 'base':297 'bash':183,422,476,842,909,945,986,1141,1176,1210 'bench':706,806 'best':232,256,679,765,880 'better':238,1359 'big':17,74,941 'break':1372 'built':1030 'built-in':1029 'call':119 'cannot':129 'capabl':282,835 'captur':416 'case':526,838 'challeng':937 'check':184,1142,1189,1413 'choos':1304 'chosen':1057 'ci/cd':848 'clarif':1074 'claud':116,846 'cleanup':1231 'clear':1286 'cli':10,1015,1400,1404 'code':12,43,56,59,117,234,319,532,576,592,694,702,839,847,864,875,904,982,1317,1370 'codebas':84,442,467,963 'codex':1025 'come':820 'command':336,414,1067,1087,1097,1187 'common':836 'complet':497,505,1276 'complex':235,692,830,1315,1374 'compon':934 'comprehens':23,58,863,903 'concern':931 'confirm':1069 'connect':1174 'context':18,27,75,82,681,942,1045,1280,1379 'continu':370,519 'control':351 'convent':973 'cost':278,638,755,1346 'cost-effici':277,754 'cost-optim':637,1345 'cpu':170,1164 'critic':98,260,626,720,1332 'cutoff':817 'debt':977 'deep':825 'deep-think':824 'default':34,107,156,301,347,405,578,588,669,989,1004,1262,1324 'detail':1178 'detect':1140 'diagnosi':1175 'diamond':809 'direct':1093 'directori':374,376,652,659,662,1297,1302 'distil':252,726 'document':85 'e':1149,1197 'edit':311,315,349,493,597,599,610,667 'effect':1384 'effici':279,756 'elo':815 'engin':241,797,1362 'enhanc':833 'ensur':1274 'entir':83,962 'error':1075 'establish':1199 'etc':849 'etim':1184 'even':1376 'except':1364 'execut':143,367,845,912,948 'exit':1088 'explicit':1298 'explor':520 'f':199,1217 'fail':1096 'failur':1080 'fastest':290,777 'featur':684 'file':63,89,97 'find':522,1048,1387 'first':1269 'fix':180 'flag':530,1103 'flagship':230,690 'flash':247,274,286,632,645,713,752,774,1328,1343 'flash-lit':285,773 'flow':1008 'focus':866 'follow':515,1012,1036,1052 'follow-up':514,1035,1051 'fulli':141 'function':1033 'gather':1073 'gemini':1,9,30,44,50,150,163,190,200,201,225,245,262,272,283,429,431,454,456,477,479,496,502,511,536,549,564,566,580,601,615,630,643,685,711,733,750,771,789,822,850,852,890,892,913,915,949,951,994,996,1014,1042,1066,1082,1086,1138,1150,1151,1213,1218,1219,1235,1256,1309,1326,1341,1354,1389,1403,1409,1415 'get':1177 'gpqa':808 'grant':1119 'grep':189,191,193,1148,1153,1155,1196,1234,1236,1238 'guid':46,677 'handl':1076 'hang':122 'high':758,779,1101,1349 'high-impact':1100 'high-volum':757,1348 'higher':793 'hung':162,181,186,1137,1144,1208,1212 'ideal':21 'impact':1102 'includ':373,378,658,661,1122,1296 'include-directori':372,657,660,1295 'indefinit':123 'info':1180 'inform':498 'initi':474 'input':698,710,723,743,762,782 'integr':936 'interact':113,158,307,365,371,409,470,575,589,663,981,992,1023,1266 'isol':388 'issu':445,873,1011 'januari':818 'key':529,683,970 'kill':194,1211,1225,1228 'knowledg':816 'l':1201 'larg':26,1279,1378 'latenc':251,717,788 'later':1407 'legaci':265,275,288,736,753,776 'leverag':1353 'like':115 'limit':451 'lite':287,775 'look':1156 'lsof':1192 'm':340,430,455,478,535,548,565,579,600,614,629,642,851,891,914,950,995 'maintain':878 'mani':96 'matur':748 'maximum':785,1320 'min':450 'minim':787 'minut':167,887,1163 'miss':933 'mode':100,106,138,155,296,346,360,386,397,421,438,463,491,528,543,556,573,587,608,622,635,648,674,747,770,859,899,922,958,990,1003,1061,1106,1244,1261 'model':216,224,231,341,342,675,678,691,1058,1307,1411 'monitor':1268 'multi':88,651 'multi-directori':650 'multi-fil':87 'multipl':62 'necessari':196 'network':173,1173,1190 'never':102,151,403,1257 'new':510,1041 'next':1070 'non':112,408,1090,1265 'non-interact':111,407,1264 'non-zero':1089 'o':1182 'one':1020 'one-shot':1019 'opportun':980 'optim':639,732,1347 'option':266,276,289,339 'output':418,700,725,745,764,784,1121,1391 'overal':967 'p':1188,1193 'partial':1125 'pattern':94,971 'pcpu':1185 'perform':271,741,861,872,901 'permiss':1114 'pid':1183,1227 'pkill':197,1215 'plan':14,64,68,906,927 'potenti':975 'practic':881 'prevent':1239 'preview':205,229,435,460,483,540,553,570,584,605,619,689,856,896,919,955,1000,1223,1313 'previous':1047 'price':766 'pro':32,204,228,244,255,264,434,459,482,539,552,569,583,604,618,688,729,735,855,895,918,954,999,1222,1312 'pro-preview':203,227,433,458,481,538,551,568,582,603,617,687,854,894,917,953,998,1221,1311 'process':20,76,164,175,182,291,778,1139,1145,1160,1168,1179,1206,1214,1275,1281,1351 'project':72 'prompt':127,222,302,364,366,368,475,665,671,1288 'prompt-interact':363 'propos':1050 'provid':131,1285 'ps':187,1146,1181,1232,1272 'qualiti':876,1322 'quick':523 'reason':41,236,693,831,1316 'recommend':423,1323 'refactor':612,979 'refer':524,1399 'relationship':92 'relev':1301 'report':1079,1393 'request':1092 'requir':25,78,329,1334,1402 'resolut':1209 'respons':1338 'restat':1055 'result':1126,1203 'resum':1032 'retri':1095 'review':13,15,57,60,65,134,320,440,465,485,533,577,593,840,865,905,907,924,983,1005 'right':1306 'roadmap':73 'run':8,165,206,383,412,1161,1270 'safeti':885,1253 'sandbox':382,385,1109 'save':1386,1397 'scalabl':930 'second':250,716,1337 'secur':444,869,1010 'select':293,343,676 'session':309,471,512,1016,1043 'set':86 'shell':114,410,1267 'shot':1021 'show':177 'singl':221 'skill':45 'skill-gemini' 'sleep':179,1171 'softwar':240,796,1361 'soon':821 'source-softaworks' 'specif':70,1226,1284 'specifi':1299 'speed':259,625,719,786,1331 'speed-crit':258,624,718,1330 'stabil':749 'start':508,1039 'stat':1186 'state':37,176,799,1169 'state-of-the-art':36,798 'step':1071 'stop':1077 'strength':1357 'strong':267,737 'structur':1287,1380,1392 'sub':249,715,1336 'sub-second':248,714,1335 'suggest':322 'summar':1127 'support':1412 'swe':705,805 'swe-bench':704,804 'symptom':160,1159 'system':488 'task':77,208,261,300,332,392,428,696,760,1248,1318,1333,1363,1375 'tcp':1198 'technic':69,976 'termin':159,308,590,984,993 'think':281,746,769,826,834 'throughput':780 'time':1339 'timeout':147,148,401,448,452,560,562,884,888,1250,1254 'tip':1277 'token':80 'tool':118,316,328,352 'topic-agent-skills' 'topic-automation' 'topic-claude' 'topic-claude-code' 'topic-coding-agent' 'topic-development' 'tpu':731 'tpu-optim':730 'troubleshoot':1136 'typic':1018 'ultra':829 'ultra-complex':828 'understand':91,965 'unless':1117 'unlik':1024 'usag':171,1165 'use':2,29,49,103,135,152,218,394,404,424,525,837,988,1062,1099,1115,1134,1241,1258,1294,1308,1325,1340 'user':5,212,500,1112 'v':192,1154,1237 'v0.16.0':1405 'verifi':1230 'version':1083,1401,1414,1416 'via':213 'vibe':701,1369 'violat':882 'volum':759,1350 'vulner':870 'wait':124 'warn':101,1123 'wc':1200 'webdev':812 'whenev':1081 'window':28,682 'without':157 'workflow':1367 'workspac':380 'wrap':145 'wrapper':402,1251 'y':354,1108 'yolo':139,323,350,355,361,398,425,439,464,534,544,547,557,561,574,613,623,628,636,641,649,654,860,900,923,959,1107,1245 'zero':1091","prices":[{"id":"7ed9fdad-9ec2-407a-91ea-df7748142d53","listingId":"65c72974-bbbf-4d9a-97d6-dddb8a0e6178","amountUsd":"0","unit":"free","nativeCurrency":null,"nativeAmount":null,"chain":null,"payTo":null,"paymentMethod":"skill-free","isPrimary":true,"details":{"org":"softaworks","category":"agent-toolkit","install_from":"skills.sh"},"createdAt":"2026-04-18T20:27:52.449Z"}],"sources":[{"listingId":"65c72974-bbbf-4d9a-97d6-dddb8a0e6178","source":"github","sourceId":"softaworks/agent-toolkit/gemini","sourceUrl":"https://github.com/softaworks/agent-toolkit/tree/main/skills/gemini","isPrimary":false,"firstSeenAt":"2026-04-18T21:54:33.407Z","lastSeenAt":"2026-05-18T18:52:57.591Z"},{"listingId":"65c72974-bbbf-4d9a-97d6-dddb8a0e6178","source":"skills_sh","sourceId":"softaworks/agent-toolkit/gemini","sourceUrl":"https://skills.sh/softaworks/agent-toolkit/gemini","isPrimary":true,"firstSeenAt":"2026-04-18T20:27:52.449Z","lastSeenAt":"2026-05-07T22:40:50.952Z"}],"details":{"listingId":"65c72974-bbbf-4d9a-97d6-dddb8a0e6178","quickStartSnippet":null,"exampleRequest":null,"exampleResponse":null,"schema":null,"openapiUrl":null,"agentsTxtUrl":null,"citations":[],"useCases":[],"bestFor":[],"notFor":[],"kindDetails":{"org":"softaworks","slug":"gemini","github":{"repo":"softaworks/agent-toolkit","stars":1847,"topics":["agent-skills","ai","automation","claude","claude-code","coding-agent","development"],"license":"mit","html_url":"https://github.com/softaworks/agent-toolkit","pushed_at":"2026-03-05T16:46:24Z","description":"A curated collection of skills for AI coding agents. Skills are packaged instructions and scripts that extend agent capabilities across development, documentation, planning, and professional workflows.","skill_md_sha":"d6f0e35147a920b13f7cf9fac4ec834527c464f5","skill_md_path":"skills/gemini/SKILL.md","default_branch":"main","skill_tree_url":"https://github.com/softaworks/agent-toolkit/tree/main/skills/gemini"},"layout":"multi","source":"github","category":"agent-toolkit","frontmatter":{"name":"gemini","description":"Use when the user asks to run Gemini CLI for code review, plan review, or big context (>200k) processing. Ideal for comprehensive analysis requiring large context windows. Uses Gemini 3 Pro by default for state-of-the-art reasoning and coding."},"skills_sh_url":"https://skills.sh/softaworks/agent-toolkit/gemini"},"updatedAt":"2026-05-18T18:52:57.591Z"}}