{"id":"6250c780-9a13-4709-9d84-b5322d5155af","shortId":"zjP8TC","kind":"skill","title":"oracle-codex","tagline":"This skill should be used when the user asks to consult Codex/GPT for planning or code review, or wants a second opinion on code. Trigger phrases include \"ask Codex\", \"consult Codex\", \"second opinion on code\", \"consult the oracle\". NOT for implementation tasks.","description":"# Codex Oracle\n\nUse OpenAI Codex CLI as a **read-only oracle** — planning, review, and analysis only. Codex provides its perspective; you synthesize and present results to the user.\n\n**Sandbox is always `read-only`**. Codex must never implement changes.\n\n## Arguments\n\nParse `$ARGUMENTS` for:\n\n- **query** — the main question or task (everything not a flag). **Required** — if empty, tell the user to provide a query and stop.\n- `--reasoning <level>` — override reasoning effort (`low`, `medium`, `high`, `xhigh`). Optional; default is `xhigh`.\n\n## Prerequisites\n\nRun the check script before any Codex invocation:\n\n```bash\nscripts/check-codex.sh\n```\n\nIf it exits non-zero, display the error and stop. Use the wrapper for all `codex exec` calls:\n\n```bash\nscripts/run-codex-exec.sh\n```\n\n## Configuration\n\n| Setting   | Default     | Override                                         |\n| --------- | ----------- | ------------------------------------------------ |\n| Model     | `gpt-5.5`   | Allowlist only (see `references/codex-flags.md`) |\n| Reasoning | `xhigh`     | `--reasoning <level>` or user prose              |\n| Sandbox   | `read-only` | Not overridable                                  |\n\n### Reasoning Effort\n\n| Complexity | Effort   | Timeout  | Criteria                             |\n| ---------- | -------- | -------- | ------------------------------------ |\n| Simple     | `low`    | 300000ms | \\<3 files, quick question            |\n| Moderate   | `medium` | 300000ms | 3–10 files, focused analysis         |\n| Complex    | `high`   | 600000ms | Multi-module, architectural thinking |\n| Maximum    | `xhigh`  | 600000ms | Full codebase, critical decisions    |\n\nFor `xhigh` tasks that may exceed 10 minutes, use `run_in_background: true` on the Bash tool and set `CODEX_OUTPUT` so you can read the output later.\n\nSee `references/codex-flags.md` for full flag documentation.\n\n## Workflow\n\n### 1. Parse and Validate\n\n1. Parse `$ARGUMENTS` for query and `--reasoning`\n2. Run `scripts/check-codex.sh` — abort on failure\n3. Use `xhigh` reasoning effort unless `--reasoning` overrides it\n\n### 2. Construct Prompt\n\nBuild a focused prompt from the user's query and any relevant context (diffs, file contents, prior conversation). Keep it direct — state what you want Codex to analyze and what kind of output you need. Do not implement; request analysis and recommendations only.\n\n### 3. Execute\n\nInvoke via the wrapper with HEREDOC. Set the Bash tool timeout per the reasoning effort table above.\n\n```bash\nEFFORT=\"<effort>\" \\\nCODEX_OUTPUT=\"/tmp/codex-${RANDOM}${RANDOM}.txt\" \\\nscripts/run-codex-exec.sh <<'EOF'\n[constructed prompt]\nEOF\n```\n\nFor `xhigh`, consider `run_in_background: true` on the Bash tool call, then read `CODEX_OUTPUT` when done.\n\n### 4. Present Results\n\nRead the output file and present with attribution:\n\n```\n## Codex Analysis\n\n[Codex output — summarize if >200 lines]\n\n---\nModel: gpt-5.5 | Reasoning: [effort level]\n```\n\nSynthesize key insights and actionable items for the user.","tags":["oracle","codex","agent","skills","paulrberg","agent-skills","ai-agents"],"capabilities":["skill","source-paulrberg","skill-oracle-codex","topic-agent-skills","topic-ai-agents"],"categories":["agent-skills"],"synonyms":[],"warnings":[],"endpointUrl":"https://skills.sh/PaulRBerg/agent-skills/oracle-codex","protocol":"skill","transport":"skills-sh","auth":{"type":"none","details":{"cli":"npx skills add PaulRBerg/agent-skills","source_repo":"https://github.com/PaulRBerg/agent-skills","install_from":"skills.sh"}},"qualityScore":"0.478","qualityRationale":"deterministic score 0.48 from registry signals: · indexed on github topic:agent-skills · 56 github stars · SKILL.md body (3,008 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":"2026-05-18T18:57:36.948Z","embedding":null,"createdAt":"2026-04-18T22:17:47.135Z","updatedAt":"2026-05-18T18:57:36.948Z","lastSeenAt":"2026-05-18T18:57:36.948Z","tsv":"'-5.5':162,393 '/tmp/codex-':345 '1':250,254 '10':196,221 '2':261,276 '200':389 '3':188,195,267,322 '300000ms':187,194 '4':372 '600000ms':202,210 'abort':264 'action':401 'allowlist':163 'alway':77 'analysi':61,199,318,384 'analyz':306 'architectur':206 'argument':86,88,256 'ask':12,31 'attribut':382 'background':226,359 'bash':133,154,230,332,341,363 'build':279 'call':153,365 'chang':85 'check':127 'cli':51 'code':19,27,38 'codebas':212 'codex':3,32,34,46,50,63,81,131,151,234,304,343,368,383,385 'codex/gpt':15 'complex':181,200 'configur':156 'consid':356 'construct':277,351 'consult':14,33,39 'content':294 'context':291 'convers':296 'criteria':184 'critic':213 'decis':214 'default':121,158 'diff':292 'direct':299 'display':141 'document':248 'done':371 'effort':115,180,182,271,338,342,395 'empti':102 'eof':350,353 'error':143 'everyth':96 'exceed':220 'exec':152 'execut':323 'exit':137 'failur':266 'file':189,197,293,378 'flag':99,247 'focus':198,281 'full':211,246 'gpt':161,392 'heredoc':329 'high':118,201 'implement':44,84,316 'includ':30 'insight':399 'invoc':132 'invok':324 'item':402 'keep':297 'key':398 'kind':309 'later':242 'level':396 'line':390 'low':116,186 'main':92 'maximum':208 'may':219 'medium':117,193 'minut':222 'model':160,391 'moder':192 'modul':205 'multi':204 'multi-modul':203 'must':82 'need':313 'never':83 'non':139 'non-zero':138 'openai':49 'opinion':25,36 'option':120 'oracl':2,41,47,57 'oracle-codex':1 'output':235,241,311,344,369,377,386 'overrid':113,159,178,274 'pars':87,251,255 'per':335 'perspect':66 'phrase':29 'plan':17,58 'prerequisit':124 'present':70,373,380 'prior':295 'prompt':278,282,352 'prose':172 'provid':64,107 'queri':90,109,258,287 'question':93,191 'quick':190 'random':346,347 'read':55,79,175,239,367,375 'read-on':54,78,174 'reason':112,114,167,169,179,260,270,273,337,394 'recommend':320 'references/codex-flags.md':166,244 'relev':290 'request':317 'requir':100 'result':71,374 'review':20,59 'run':125,224,262,357 'sandbox':75,173 'script':128 'scripts/check-codex.sh':134,263 'scripts/run-codex-exec.sh':155,349 'second':24,35 'see':165,243 'set':157,233,330 'simpl':185 'skill':5 'skill-oracle-codex' 'source-paulrberg' 'state':300 'stop':111,145 'summar':387 'synthes':68,397 'tabl':339 'task':45,95,217 'tell':103 'think':207 'timeout':183,334 'tool':231,333,364 'topic-agent-skills' 'topic-ai-agents' 'trigger':28 'true':227,360 'txt':348 'unless':272 'use':8,48,146,223,268 'user':11,74,105,171,285,405 'valid':253 'via':325 'want':22,303 'workflow':249 'wrapper':148,327 'xhigh':119,123,168,209,216,269,355 'zero':140","prices":[{"id":"44e54559-8f46-4254-863a-5c1435d6ef27","listingId":"6250c780-9a13-4709-9d84-b5322d5155af","amountUsd":"0","unit":"free","nativeCurrency":null,"nativeAmount":null,"chain":null,"payTo":null,"paymentMethod":"skill-free","isPrimary":true,"details":{"org":"PaulRBerg","category":"agent-skills","install_from":"skills.sh"},"createdAt":"2026-04-18T22:17:47.135Z"}],"sources":[{"listingId":"6250c780-9a13-4709-9d84-b5322d5155af","source":"github","sourceId":"PaulRBerg/agent-skills/oracle-codex","sourceUrl":"https://github.com/PaulRBerg/agent-skills/tree/main/skills/oracle-codex","isPrimary":false,"firstSeenAt":"2026-04-18T22:17:47.135Z","lastSeenAt":"2026-05-18T18:57:36.948Z"}],"details":{"listingId":"6250c780-9a13-4709-9d84-b5322d5155af","quickStartSnippet":null,"exampleRequest":null,"exampleResponse":null,"schema":null,"openapiUrl":null,"agentsTxtUrl":null,"citations":[],"useCases":[],"bestFor":[],"notFor":[],"kindDetails":{"org":"PaulRBerg","slug":"oracle-codex","github":{"repo":"PaulRBerg/agent-skills","stars":56,"topics":["agent-skills","ai-agents"],"license":"mit","html_url":"https://github.com/PaulRBerg/agent-skills","pushed_at":"2026-05-17T10:33:19Z","description":"PRB's collection of agent skills","skill_md_sha":"da722ab8e78df4637577ba0faaf7bb981b720714","skill_md_path":"skills/oracle-codex/SKILL.md","default_branch":"main","skill_tree_url":"https://github.com/PaulRBerg/agent-skills/tree/main/skills/oracle-codex"},"layout":"multi","source":"github","category":"agent-skills","frontmatter":{"name":"oracle-codex","description":"This skill should be used when the user asks to consult Codex/GPT for planning or code review, or wants a second opinion on code. Trigger phrases include \"ask Codex\", \"consult Codex\", \"second opinion on code\", \"consult the oracle\". NOT for implementation tasks."},"skills_sh_url":"https://skills.sh/PaulRBerg/agent-skills/oracle-codex"},"updatedAt":"2026-05-18T18:57:36.948Z"}}