{"id":"0e491ddf-3cb0-42a4-acb2-9dbd1ca3b0f1","shortId":"An3kRd","kind":"skill","title":"codex-agent","tagline":"MANDATORY for code review - must use Codex CLI for all code reviews, then apply fixes based on Codex feedback. Also use for cross-verification, debugging, and getting alternative implementations.","description":"# Codex Agent Collaboration Skill\n\nThis skill enables Claude Code to collaborate with OpenAI's Codex CLI agent.\n\n## MANDATORY: Code Review Workflow\n\n**IMPORTANT**: When performing code review, you MUST follow this workflow:\n\n### Step 1: Call Codex for Review\n\n```bash\ncodex exec -C <project_path> -s read-only -o /tmp/codex-review.md \\\n  \"Review the code in <file_or_directory>. Check for:\n   - Security vulnerabilities\n   - Performance issues\n   - Code quality and best practices\n   - Potential bugs and edge cases\n   - Naming and readability\n   Provide specific, actionable feedback with file paths and line numbers.\"\n```\n\n### Step 2: Read Codex Feedback\n\n```bash\ncat /tmp/codex-review.md\n```\n\n### Step 3: Apply Fixes Based on Codex Feedback\n\nFor each issue identified by Codex:\n1. Read the relevant file\n2. Apply the fix using Edit tool\n3. Verify the fix addresses Codex's concern\n\n### Step 4: Re-verify with Codex (Optional)\n\n```bash\ncodex exec -C <project_path> -s read-only \\\n  \"Verify the fixes applied to <files>. Confirm issues are resolved.\"\n```\n\n## Workflow Examples\n\n### Example 1: Review and Fix a Single File\n\n```bash\n# Step 1: Get Codex review\ncodex exec -C /project -s read-only -o /tmp/codex-review.md \\\n  \"Review src/auth/login.ts for security vulnerabilities and code quality issues. Provide specific line numbers and fixes.\"\n\n# Step 2: Read the feedback\ncat /tmp/codex-review.md\n```\n\nThen Claude reads the feedback, applies fixes with Edit tool, and optionally re-verifies.\n\n### Example 2: Review Recent Changes\n\n```bash\n# Get diff of recent changes\ngit diff HEAD~1 > /tmp/recent-changes.diff\n\n# Step 1: Have Codex review the diff\ncodex exec -C /project -s read-only -o /tmp/codex-review.md \\\n  \"Review the changes in the last commit. Check for bugs, security issues, and improvements needed.\"\n\n# Step 2: Read and apply fixes\ncat /tmp/codex-review.md\n```\n\n### Example 3: Full Project Review\n\n```bash\n# Step 1: Comprehensive review\ncodex exec -C /project -s read-only -o /tmp/codex-review.md \\\n  \"Perform a comprehensive code review of src/. Focus on:\n   1. Security vulnerabilities (OWASP Top 10)\n   2. Error handling patterns\n   3. Performance bottlenecks\n   4. Code duplication\n   Prioritize issues by severity (critical/high/medium/low).\"\n\n# Step 2: Read prioritized feedback\ncat /tmp/codex-review.md\n```\n\n## Review Request Format\n\nWhen asking Codex for review, include:\n\n```\nReview <target_files_or_directory>.\n\nContext:\n- Project type: <TypeScript/Python/etc>\n- Framework: <Express/React/etc>\n- Focus areas: <security/performance/quality>\n\nCheck for:\n1. Security vulnerabilities\n2. Performance issues\n3. Error handling\n4. Code quality\n5. Edge cases\n\nOutput format:\nFor each issue:\n- File: <path>\n- Line: <number>\n- Severity: critical/high/medium/low\n- Issue: <description>\n- Fix: <specific code change>\n```\n\n## Applying Fixes\n\nAfter receiving Codex feedback, apply fixes systematically:\n\n1. **Parse the review** - Extract each issue with file, line, severity\n2. **Prioritize** - Fix critical/high issues first\n3. **Read file** - Use Read tool to see current code\n4. **Apply fix** - Use Edit tool with precise old_string/new_string\n5. **Track progress** - Mark each issue as fixed\n\n## Prerequisites\n\nCodex CLI must be installed and authenticated:\n\n```bash\n# Install via npm\nnpm install -g @openai/codex\n\n# Or via Homebrew (macOS)\nbrew install --cask codex\n\n# Authenticate\ncodex login\n```\n\n## Command Reference\n\n### Basic Command Pattern\n\n```bash\ncodex exec [options] \"<task_description>\"\n```\n\n### Core Options\n\n| Option | Description |\n|--------|-------------|\n| `\"<task>\"` | Task description (positional, must be quoted) |\n| `-C <dir>` | Working directory (use absolute path) |\n| `-s read-only` | Read-only sandbox (use for reviews) |\n| `-o <path>` | Save output to file |\n| `--json` | Output as JSON Lines |\n\n### AI-to-AI Communication\n\nWhen communicating with Codex, PRIORITIZE ACCURACY AND PRECISION:\n- Use structured data and exact technical terms\n- Provide full file paths and precise details\n- Include relevant context from the current codebase\n- NO conversational formatting needed\n\n## Other Use Cases\n\n### Cross-Verification (after Claude implements)\n\n```bash\ncodex exec -C /project -s read-only \\\n  \"Verify the implementation in src/feature/. Check correctness and edge cases.\"\n```\n\n### Get Alternative Implementation\n\n```bash\ncodex exec -C /project -s read-only -o /tmp/alternative.md \\\n  \"Propose an alternative implementation for the caching in src/cache/manager.ts\"\n```\n\n### Debugging Assistance\n\n```bash\ncodex exec -C /project -s read-only \\\n  \"Debug: tests in tests/auth.test.ts failing with timeout. Analyze root cause.\"\n```\n\n## Session Management\n\nFor multi-turn reviews:\n\n```bash\n# Initial review\ncodex exec -C /project -s read-only \"Review src/api/ for security issues\"\n# Note session ID from output\n\n# Follow-up after fixes\ncodex exec resume <session_id> \"I've applied the fixes. Please re-verify.\"\n```\n\n## Troubleshooting\n\n### Authentication Issues\n\n```bash\ncodex logout\ncodex login\n```\n\n### Check Installation\n\n```bash\ncodex --version\nwhich codex\n```\n\n## See Also\n\n- [sandbox-modes.md](sandbox-modes.md) - Sandbox security levels\n- [examples.md](examples.md) - More usage examples\n- [advanced.md](advanced.md) - Advanced configuration","tags":["codex","agent","claude","arsenal","majiayu000","agent-skills","ai-agents","ai-coding-assistant","automation","claude-code","code-review","developer-tools"],"capabilities":["skill","source-majiayu000","skill-codex-agent","topic-agent-skills","topic-ai-agents","topic-ai-coding-assistant","topic-automation","topic-claude","topic-claude-code","topic-code-review","topic-developer-tools","topic-devops","topic-productivity","topic-prompt-engineering","topic-python"],"categories":["claude-arsenal"],"synonyms":[],"warnings":[],"endpointUrl":"https://skills.sh/majiayu000/claude-arsenal/codex-agent","protocol":"skill","transport":"skills-sh","auth":{"type":"none","details":{"cli":"npx skills add majiayu000/claude-arsenal","source_repo":"https://github.com/majiayu000/claude-arsenal","install_from":"skills.sh"}},"qualityScore":"0.464","qualityRationale":"deterministic score 0.46 from registry signals: · indexed on github topic:agent-skills · 29 github stars · SKILL.md body (5,281 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":"2026-05-01T07:01:13.103Z","embedding":null,"createdAt":"2026-04-18T22:24:03.953Z","updatedAt":"2026-05-01T07:01:13.103Z","lastSeenAt":"2026-05-01T07:01:13.103Z","tsv":"'/project':200,270,313,582,604,626,654 '/tmp/alternative.md':610 '/tmp/codex-review.md':80,121,206,228,276,299,319,356 '/tmp/recent-changes.diff':259 '1':66,136,184,193,261,307,329,378,413 '10':334 '2':115,141,223,245,293,335,351,381,424 '3':123,148,301,339,384,430 '4':157,342,387,440 '5':390,450 'absolut':508 'accuraci':541 'action':106 'address':152 'advanc':715 'advanced.md':713,714 'agent':3,35,50 'ai':532,534 'ai-to-ai':531 'also':23,702 'altern':32,598,613 'analyz':638 'appli':17,124,142,175,234,296,404,410,441,679 'area':374 'ask':361 'assist':621 'authent':465,482,687 'base':19,126 'bash':71,119,164,191,249,305,466,490,578,600,622,648,689,696 'basic':487 'best':94 'bottleneck':341 'brew':478 'bug':97,286 'c':74,167,199,269,312,504,581,603,625,653 'cach':617 'call':67 'case':100,392,571,596 'cask':480 'cat':120,227,298,355 'caus':640 'chang':248,254,279 'check':85,284,376,592,694 'claud':41,230,576 'cli':11,49,460 'code':6,14,42,52,58,83,91,213,323,343,388,439 'codebas':564 'codex':2,10,21,34,48,68,72,117,128,135,153,162,165,195,197,263,267,310,362,408,459,481,483,491,539,579,601,623,651,674,690,692,697,700 'codex-ag':1 'collabor':36,44 'command':485,488 'commit':283 'communic':535,537 'comprehens':308,322 'concern':155 'configur':716 'confirm':177 'context':367,560 'convers':566 'core':494 'correct':593 'critical/high':427 'critical/high/medium/low':349,401 'cross':27,573 'cross-verif':26,572 'current':438,563 'data':546 'debug':29,620,631 'descript':497,499 'detail':557 'diff':251,256,266 'directori':506 'duplic':344 'edg':99,391,595 'edit':146,237,444 'enabl':40 'error':336,385 'exact':548 'exampl':182,183,244,300,712 'examples.md':708,709 'exec':73,166,198,268,311,492,580,602,624,652,675 'express/react/etc':372 'extract':417 'fail':635 'feedback':22,107,118,129,226,233,354,409 'file':109,140,190,398,421,432,525,553 'first':429 'fix':18,125,144,151,174,187,221,235,297,403,405,411,426,442,457,673,681 'focus':327,373 'follow':62,670 'follow-up':669 'format':359,394,567 'framework':371 'full':302,552 'g':472 'get':31,194,250,597 'git':255 'handl':337,386 'head':257 'homebrew':476 'id':666 'identifi':133 'implement':33,577,589,599,614 'import':55 'improv':290 'includ':365,558 'initi':649 'instal':463,467,471,479,695 'issu':90,132,178,215,288,346,383,397,402,419,428,455,663,688 'json':526,529 'last':282 'level':707 'line':112,218,399,422,530 'login':484,693 'logout':691 'maco':477 'manag':642 'mandatori':4,51 'mark':453 'multi':645 'multi-turn':644 'must':8,61,461,501 'name':101 'need':291,568 'note':664 'npm':469,470 'number':113,219 'o':79,205,275,318,521,609 'old':448 'openai':46 'openai/codex':473 'option':163,240,493,495,496 'output':393,523,527,668 'owasp':332 'pars':414 'path':110,509,554 'pattern':338,489 'perform':57,89,320,340,382 'pleas':682 'posit':500 'potenti':96 'practic':95 'precis':447,543,556 'prerequisit':458 'priorit':345,353,425,540 'progress':452 'project':303,368 'propos':611 'provid':104,216,551 'qualiti':92,214,389 'quot':503 're':159,242,684 're-verifi':158,241,683 'read':77,116,137,170,203,224,231,273,294,316,352,431,434,512,515,585,607,629,657 'read-on':76,169,202,272,315,511,514,584,606,628,656 'readabl':103 'receiv':407 'recent':247,253 'refer':486 'relev':139,559 'request':358 'resolv':180 'resum':676 'review':7,15,53,59,70,81,185,196,207,246,264,277,304,309,324,357,364,366,416,520,647,650,659 'root':639 'sandbox':517,705 'sandbox-modes.md':703,704 'save':522 'secur':87,210,287,330,379,662,706 'security/performance/quality':375 'see':437,701 'session':641,665 'sever':348,400,423 'singl':189 'skill':37,39 'skill-codex-agent' 'source-majiayu000' 'specif':105,217 'src':326 'src/api':660 'src/auth/login.ts':208 'src/cache/manager.ts':619 'src/feature':591 'step':65,114,122,156,192,222,260,292,306,350 'string/new_string':449 'structur':545 'systemat':412 'task':498 'technic':549 'term':550 'test':632 'tests/auth.test.ts':634 'timeout':637 'tool':147,238,435,445 'top':333 'topic-agent-skills' 'topic-ai-agents' 'topic-ai-coding-assistant' 'topic-automation' 'topic-claude' 'topic-claude-code' 'topic-code-review' 'topic-developer-tools' 'topic-devops' 'topic-productivity' 'topic-prompt-engineering' 'topic-python' 'track':451 'troubleshoot':686 'turn':646 'type':369 'typescript/python/etc':370 'usag':711 'use':9,24,145,433,443,507,518,544,570 've':678 'verif':28,574 'verifi':149,160,172,243,587,685 'version':698 'via':468,475 'vulner':88,211,331,380 'work':505 'workflow':54,64,181 '~1':258","prices":[{"id":"b877005d-46dd-4a74-971e-0c1807536fdc","listingId":"0e491ddf-3cb0-42a4-acb2-9dbd1ca3b0f1","amountUsd":"0","unit":"free","nativeCurrency":null,"nativeAmount":null,"chain":null,"payTo":null,"paymentMethod":"skill-free","isPrimary":true,"details":{"org":"majiayu000","category":"claude-arsenal","install_from":"skills.sh"},"createdAt":"2026-04-18T22:24:03.953Z"}],"sources":[{"listingId":"0e491ddf-3cb0-42a4-acb2-9dbd1ca3b0f1","source":"github","sourceId":"majiayu000/claude-arsenal/codex-agent","sourceUrl":"https://github.com/majiayu000/claude-arsenal/tree/main/skills/codex-agent","isPrimary":false,"firstSeenAt":"2026-04-18T22:24:03.953Z","lastSeenAt":"2026-05-01T07:01:13.103Z"}],"details":{"listingId":"0e491ddf-3cb0-42a4-acb2-9dbd1ca3b0f1","quickStartSnippet":null,"exampleRequest":null,"exampleResponse":null,"schema":null,"openapiUrl":null,"agentsTxtUrl":null,"citations":[],"useCases":[],"bestFor":[],"notFor":[],"kindDetails":{"org":"majiayu000","slug":"codex-agent","github":{"repo":"majiayu000/claude-arsenal","stars":29,"topics":["agent-skills","ai-agents","ai-coding-assistant","automation","claude","claude-code","code-review","developer-tools","devops","productivity","prompt-engineering","python","software-development","typescript","workflows"],"license":"mit","html_url":"https://github.com/majiayu000/claude-arsenal","pushed_at":"2026-04-29T04:12:22Z","description":"52 production-ready Claude Code skills and 7 specialized agents for software development, DevOps, product workflows, and automation.","skill_md_sha":"1a475e92c5aac1a6c8dee89367b8cb163fa87536","skill_md_path":"skills/codex-agent/SKILL.md","default_branch":"main","skill_tree_url":"https://github.com/majiayu000/claude-arsenal/tree/main/skills/codex-agent"},"layout":"multi","source":"github","category":"claude-arsenal","frontmatter":{"name":"codex-agent","description":"MANDATORY for code review - must use Codex CLI for all code reviews, then apply fixes based on Codex feedback. Also use for cross-verification, debugging, and getting alternative implementations."},"skills_sh_url":"https://skills.sh/majiayu000/claude-arsenal/codex-agent"},"updatedAt":"2026-05-01T07:01:13.103Z"}}