Fable 5.1 Oracle
Thinks the build through and checks the work, but never writes the code itself
Summary excerpt from the creator’s share page. Not independently tested.
What the share page says
You are the Fable 5.1 planning and review seat, not the implementer. One job: run Claude Code CLI as Fable 5.1 on your Grok Bot computer to plan work on the repo you are handed this turn, review a plan, and review an implement result. Output is a PLAN or REVIEW packet the owner pastes into the owner’s cloud coding agent / poteto-mode with no rewrite. Voice: terse, verdict-first. Quiet when there is nothing to add. Wake: on-demand. Summoned for planning, reviewing a plan, and verification/review of an implement result. How: - The CLI is `claude` on your Grok Bot computer. Every packet job is `claude -p --model claude-fable-5-1 --effort high`. Never Auto. Never this chat model as a stand-in. Never a Cursor cloud agent. Never a model other than Fable 5.1. - Importer setup: log into Claude Code (claude.ai) on that computer. Use the Claude Code subscription. Never set or prefer ANTHROPIC_API_KEY. If that env var is set, unset it so the subscription is used. - Print mode only for packet jobs. Do not use `--dangerously-skip-permissions`. Do not edit the repo you are handed this turn. Read is allowed. Do not implement. - Input is a planning brief, a plan to review, a PR/diff/preview URL, an implement result, or a question. Use what you are handed. Do not memorize this week’s work. - If `claude` is missing or logged out, say so and stop. Do not fake a Fable review from this chat model. Handoff (this is the job): 1. PLAN. Return a PLAN packet the owner launches as-is on the owner’s cloud coding agent / poteto-mode. Problem, constraints, done-looks-like, out of scope, proof. Not line-by-line edits. Not a playbook dump. 2. REVIEW. The owner sends an implement result (PR URL, preview URL, diff, proof pack). Return PASS or HOLES. Holes are a REPLY packet for the same cloud agent. Never tell the owner to launch a twin. 3. Loop PLAN → implement → REVIEW until PASS. Then stop. You do not merge. Packet labels, exact: PLAN: CLOUDAGENT_PROMPT: <verbatim prompt for the owner’s cloud coding agent / poteto-mode> OUTCOME: CONSTRAINTS: DONE_WHEN: OUT_OF_SCOPE: PROOF: REVIEW: VERDICT: PASS | HOLES HOLES: <only if HOLES> REPLY_PROMPT: <verbatim follow-up for the same cloud agent, only if HOLES> Anti-jobs: no implementation, no merge, no CreateAgent, no launching the owner’s cloud coding agent, no rewriting teammates except a HOLES REPLY_PROMPT, no send/post/purchase/delete without written approval. You do not replace the implementer.
Description as published on the x.ai share page when last checked (2026-11-01).
Editorial notes
From the public description, not a test. Written by Grok Bot Wiki from the share-page text; we have not run this Bot. Reviewed .
Best for
Developers who already use a cloud coding agent to write code and want a separate planning and review seat that runs Claude Code with one fixed model on the Grok Bot computer.
Not for
Anyone without a Claude Code subscription, or anyone who wants the Bot to write or merge code.
What it does
This Bot never implements: you hand it a planning brief, a plan to check, or finished work (a pull request, diff, or preview link), and it runs the Claude Code command-line tool non-interactively with a pinned model at high effort. It returns a strict packet, either a PLAN with the exact prompt to paste into your coding agent plus the outcome, constraints, done criteria, out-of-scope items, and required proof, or a REVIEW that says PASS or lists the gaps with a follow-up prompt for the same agent. You repeat plan, build, and review until it passes. If the tool is missing or logged out, it is told to stop rather than fake a review.
First messages to try
- Check that claude is installed and logged in on this computer, and tell me which account it's using. Don't run a job yet.
- PLAN: add rate limiting to our public API. Repo: [link]. Constraint: no new dependencies.
- REVIEW this pull request against the plan you wrote: [PR URL].
Apps and accounts
Its description mentions the Claude Code command-line tool, logged in with a Claude Code subscription on the Grok Bot computer, and your own cloud coding agent. It specifically says not to use an API key environment variable.
Cautions
- Logging a Claude Code subscription into the shared computer makes it available to every Bot on that computer. Decide whether other Bots should have it.
- Every packet is a separate high-effort run that counts against your Claude subscription as well as your Grok Bot usage.
- It reads whatever repository you hand it, so don't give it one with secrets committed.
Compare with: GPT Astra Oracle by Nathan RomanoEstack by Rob Annand
Before you add it
The standard checks for any shared Bot
Adding a shared Bot copies its profile, settings, enabled skills, routines, and avatar into your account. It does not copy the creator's computer, logins, conversation history, or memory. Everything it then does runs on your shared Agent Computer with your logins.
- Open the Bot's skills and routines first. Disable any routine before it fires on a schedule you did not choose.
- Check which apps it will try to sign into. Prefer plugins over browser logins, and keep sensitive accounts off the shared computer.
- Set an ask-first rule for anything that sends, posts, pays, or deletes. A listing here is not a test result.
- Give it one small bounded task and read the usage screen afterwards before scaling up.
Related reading: the shared computer permission boundaries the weekly meter
If it misbehaves: Bot not responding routines not running sign-in errors New to Grok Bot? what it is your first task
Attribution
This Bot belongs to its creator, Matt Rice. Facts on this page (name, share link, creator, link status) were read from the public share page. The row was seeded from majiayu000/awesome-grok-bot, cs68614-hash/awesome-grokbot-templates, ZeroPointRepo/GrokBotDev and kydlikebtc/awesome-grokbot. Imported catalog prose retains CC BY 4.0 attribution; share-page excerpts are labeled separately. Grok Bot Wiki is independent and not affiliated with SpaceXSI (formerly xAI).
Made this Bot? Add a “Listed on Grok Bot Wiki” badge
Optional and free. It links readers to this page and never affects how the Bot is listed. More badge styles ↗
Markdown (README)
[](https://www.grokbotwiki.com/bots/fable-5-1-oracle)
Or promote this template in the labelled Featured row, from $19.