Snuz Baby Sleeping Bag with Nappy Zip, Rainbow 2.5 Tog

🧿 oracle 0.8.5 — One slug to gather them all. Prompt is required. Provide it via --prompt "<text>" or positional [prompt]. Oracle CLI v0.8.5 — Prompt + files required — GPT-5.2 Pro/GPT-5.2 for tough questions with code/file context. Usage: oracle [options] [command] [prompt] One-shot GPT-5.2 Pro / GPT-5.2 / GPT-5.1 Codex tool for hard questions that benefit from large file context and server-side search. Arguments: prompt Prompt text (shorthand for --prompt). Options: -V, --version output the version number -p, --prompt <text> User prompt to send to the model. -f, --file <paths...> Files/directories or glob patterns to attach (prefix with !pattern to exclude). Files larger than 1 MB are rejected automatically. (default: []) --copy-markdown Copy the assembled markdown bundle to the clipboard; pair with --render to print it too. (default: false) -s, --slug <words> Custom session slug (3-5 words). -m, --model <model> Model to target (gpt-5.2-pro default; also supports gpt-5.1-pro alias). Also gpt-5-pro, gpt-5.1, gpt-5.1-codex API-only, gpt-5.2, gpt-5.2-instant, gpt-5.2-pro, gemini-3-pro, claude-4.5-sonnet, claude-4.1-opus, or ChatGPT labels like "5.2 Thinking" for browser runs). --models <models> Comma-separated API model list to query in parallel (e.g., "gpt-5.2-pro,gemini-3-pro"). (default: []) -e, --engine <mode> Execution engine (api | browser). Browser engine: GPT models automate ChatGPT; Gemini models use a cookie-based client for gemini.google.com. If omitted, oracle picks api when OPENAI_API_KEY is set, otherwise browser. (choices: "api", "browser") --files-report Show token usage per attached file (also prints automatically when files exceed the token budget). (default: false) -v, --verbose Enable verbose logging for all operations. (default: false) --notify Desktop notification when a session finishes (default on unless CI/SSH). --no-notify Disable desktop notifications. --notify-sound Play a notification sound on completion (default off). --no-notify-sound Disable notification sounds. --timeout <seconds|auto> Overall timeout before aborting the API call (auto = 60m for gpt-5.2-pro, 120s otherwise). (default: "auto") --background Use Responses API background mode (create + retrieve) for API runs. --no-background Disable Responses API background mode. --http-timeout <ms|s|m|h> HTTP client timeout for API requests (default 20m). --zombie-timeout <ms|s|m|h> Override stale-session cutoff used by `oracle status` (default 60m). --zombie-last-activity Base stale-session detection on last log activity instead of start time. (default: false) --dry-run [mode] Preview without calling the model (summary | json | full). (choices: "summary", "json", "full", default: false, preset: "summary") --render-markdown Print the assembled markdown bundle for prompt + files and exit; pair with --copy to put it on the clipboard. (default: false) --render Alias for --render-markdown. (default: false) --render-plain Render markdown without ANSI/highlighting (use plain text even in a TTY). (default: false) --write-output <path> Write only the final assistant message to this file (overwrites; multi-model appends .<model> before the extension). --verbose-render Show render/TTY diagnostics when replaying sessions. (default: false) --base-url <url> Override the OpenAI-compatible base URL for API runs (e.g. LiteLLM proxy endpoint). --azure-endpoint <url> Azure OpenAI Endpoint (e.g. https://resource.openai.azure.com/). --azure-deployment <name> Azure OpenAI Deployment Name. --azure-api-version <version> Azure OpenAI API Version. --browser-cookie-path <path> Explicit Chrome/Chromium cookie DB path for session reuse. --chatgpt-url <url> Override the ChatGPT web URL (e.g., workspace/folder like https://chatgpt.com/g/.../project; default https://chatgpt.com/). --browser-port <port> Use a fixed Chrome DevTools port (helpful on WSL firewalls). --browser-model-strategy <mode> ChatGPT model picker strategy: select (default) switches to the requested model, current keeps the active model, ignore skips the picker entirely. (choices: "select", "current", "ignore") --browser-attachments <mode> How to deliver --file inputs in browser mode: auto (default) pastes inline up to ~60k chars then uploads; never always paste inline; always always upload. (choices: "auto", "never", "always", default: "auto") --remote-chrome <host:port> Connect to remote Chrome DevTools Protocol (e.g., 192.168.1.10:9222 or [2001:db8::1]:9222 for IPv6). --remote-host <host:port> Delegate browser runs to a remote `oracle serve` instance. --remote-token <token> Access token for the remote `oracle serve` instance. --browser-inline-files Alias for --browser-attachments never (force pasting file contents inline). (default: false) --browser-bundle-files Bundle all attachments into a single archive before uploading. (default: false) --youtube <url> YouTube video URL to analyze (Gemini web/cookie mode only; uses your signed-in Chrome cookies for gemini.google.com). --generate-image <file> Generate image and save to file (Gemini web/cookie mode only; requires gemini.google.com Chrome cookies). --edit-image <file> Edit existing image (use with --output, Gemini web/cookie mode only). --output <file> Output file path for image operations (Gemini web/cookie mode only). --aspect <ratio> Aspect ratio for image generation: 16:9, 1:1, 4:3, 3:4 (Gemini web/cookie mode only). --gemini-show-thoughts Display Gemini thinking process (Gemini web/cookie mode only). (default: false) --retain-hours <hours> Prune stored sessions older than this many hours before running (set 0 to disable). --force Force start a new session even if an identical prompt is already running. (default: false) --debug-help Show the advanced/debug option set and exit. (default: false) --heartbeat <seconds> Emit periodic in-progress updates (0 to disable). (default: 30) --wait -h, --help display help for command Commands: serve [options] Run Oracle browser automation as a remote service for other machines. bridge Bridge a Windows-hosted ChatGPT session to Linux clients. tui Launch the interactive terminal UI for humans (no automation). session [options] [id] Attach to a stored session or list recent sessions when no ID is provided. status [options] [id] List recent sessions (24h window by default) or attach to a session when an ID is provided. Tips • Required: always pass a prompt AND --file … (directories/globs are fine); Oracle cannot see your project otherwise. • Attach lots of source (whole directories beat single files) and keep total input under ~196k tokens. • Oracle starts empty—open with a short project briefing (stack, services, build steps), spell out the question and prior attempts, and why it matters; the more explanation and context you provide, the better the response will be. • Spell out the project + platform + version requirements (repo name, target OS/toolchain versions, API dependencies) so Oracle doesn’t guess defaults. • When comparing multiple repos/files, spell out each repo + path + role (e.g., “Project A SettingsView → apps/project-a/Sources/SettingsView.swift; Project B SettingsView → ../project-b/mac/...”) so the model knows exactly which file is which. • Best results: 6–30 sentences plus key source files; very short prompts often yield generic answers. • Oracle is one-shot: it does not remember prior runs, so start fresh each time with full context. • Run --files-report to inspect token spend before hitting the API. • Non-preview runs spawn detached sessions (especially gpt-5.2-pro API). If the CLI times out, do not re-run — reattach with oracle session <slug> to resume/inspect the existing run. • Set a memorable 3–5 word slug via --slug "<words>" to keep session IDs tidy. • Finished sessions auto-hide preamble logs when reattached; raw timestamps remain in the saved log file. • Need hidden flags? Run oracle --help --verbose to list search/token/browser overrides. • If any Oracle session is already running, do not start new API runs. Attach to the existing browser session instead; only trigger API calls when you explicitly mean to. • Duplicate prompt guard: if the same prompt is already running, new runs are blocked unless you pass --force—prefer reattaching instead of spawning duplicates. Examples oracle --render --copy --prompt "Review the TS data layer for schema drift" --file "src/**/*.ts,*/*.test.ts" Build the bundle, print it, and copy it for manual paste into ChatGPT. oracle --prompt "Cross-check the data layer assumptions" --models gpt-5.2-pro,gemini-3-pro --file "src/**/*.ts" Run multiple API models in one go and aggregate cost/usage. oracle status --hours 72 --limit 50 Show sessions from the last 72h (capped at 50 entries). oracle session <sessionId> Attach to a running/completed session and stream the saved transcript. oracle --prompt "Ship review" --slug "release-readiness-audit" Encourage the model to hand you a 3–5 word slug and pass it along with --slug. oracle --prompt "Tabs frozen: compare Project A SettingsView (apps/project-a/Sources/SettingsView.swift) vs Project B SettingsView (../project-b/mac/App/Presentation/Views/SettingsView.swift)" --file apps/project-a/Sources/SettingsView.swift --file ../project-b/mac/App/Presentation/Views/SettingsView.swift Spell out what each attached file is (repo + path + role) before asking for comparisons so the model knows exactly what it is reading. Examples: # Quick API run with two files oracle --prompt "Summarize the risk register" --file docs/risk-register.md docs/risk-matrix.md # Browser run (no API key) + globbed TypeScript sources, excluding tests oracle --engine browser --prompt "Review the TS data layer" \ --file "src/**/*.ts" --file "!src/**/*.test.ts" # Build, print, and copy a markdown bundle (semi-manual) oracle --render --copy -p "Review the TS data layer" --file "src/**/*.ts" --file "!src/**/*.test.ts"