this is rare f*cking gold for Claude Code

give Claude Opus 5.5 its own engineering team Sonnet 5.5 handles the legwork Fable 5.1 checks the decisions that matter

plan on high → delegate on medium bring in Fable before a mistake gets expensive

keep your Opus tokens for the hard calls

• give each model a clear job

Opus 5.5, high - plans and writes code Sonnet 5.5, medium - runs the support crew explorer - reads the codebase worker - edits and runs tests researcher - fetches the docs Fable 5.1 - reviews the full session /advisor fable - adds that second opinion

• ask Fable where mistakes compound → before a plan: is this the right approach? → when an error repeats: are we looking elsewhere? → before “done”: what escaped the checks?

JEV engineering handles the smaller calls:

which file to open which tool to use whether to retry or stop bounded judgments in under half a second

the bigger models spend their attention on decisions that deserve it

paste this into Claude Code ↓

“prepare a revised Claude Code setup

1. Inspect ~/.claude/agents and .claude/agents Reuse agents suited to explorer, worker and researcher; draft only missing roles

Use model: sonnet and effort: medium List agents pinned to a different model and preserve their configuration

2. Propose these changes in ~/.claude/settings.json: effortLevel: high advisorModel: fable

3. Report advisor blockers: CLAUDE_CODE_DISABLE_ADVISOR_TOOL DISABLE_TELEMETRY anything blocking feature-flag fetching

Also report CLAUDE_CODE_EFFORT_LEVEL Leave these settings untouched

4. Propose adding this rule to ~/.claude/CLAUDE.md: consult the advisor before a major plan, when the same error repeats, and before declaring a long task complete

Show every proposed change as a diff Wait for my ‘go’ before making any edits“

↳ https://code.claude.com/docs/en/advisor

Escalate hard decisions with the advisor tool - Claude Code Docs

Source: https://code.claude.com/docs/en/advisor The advisor tool lets Claude consult a second, typically stronger model at key moments during a task, such as before committing to an approach, when stuck on a recurring error, or before declaring a task complete. The advisor receives the full conversation, including every tool call and result, and returns guidance that Claude applies before continuing.The advisor runs server-side on Anthropic’s infrastructure as aserver tool, available to both subscription and API-billed accounts. You choose which model acts as the advisor, and Claude decides when to call it.This page covers how to enable the advisor, which model pairings are accepted, what Claude shows during a consultation, and how advisor usage is billed.

When to use the advisor

The advisor fits long, multi-step tasks where most turns are routine but plan quality determines the outcome. Examples include large refactors, debugging sessions where an error keeps recurring, and tasks you want independently checked before Claude declares them done.It adds less value on short tasks where there is little to plan, or on work where every turn needs the strongest model. For those,switch the main modelinstead, or seehow the advisor compares with opusplan and subagentsfor other ways to get a second opinion.

Enable the advisor

You can set the advisor model in three ways:

• /advisorcommand: set or change the advisor mid-session and save it as your default
• advisorModelsetting: configure a persistent default in yoursettings file
• \-\-advisorflag: set the advisor for a single session at launch

Each of these enables the advisor for sessions whose main modelsupports it. After the session starts, Claude Code shows anAdvisor Tool \(experimental\) is on and may use more tokens · /advisornotification. To stop using the advisor, seeTurn the advisor off.On some plans, Fable as the advisor also needs your one-timeconsent to bill Fable usage to usage credits. For what happens before you have given that consent, seeFable advisor and usage credits.

Use the/advisorcommand

Run/advisorwithout arguments to open a picker listing the available advisor models, or pass the model directly:

The command confirms withAdvisor set tofollowed by the advisor model name. Your selection is saved toadvisorModelin your user settings and persists across sessions, except in the cases theadvisorModelentrylists as applying to the current session only.The command also works where there is no terminal picker: innon-interactive modewith\-p, in the Agent SDK, in the desktop app, and overRemote Control. This requires Claude Code v2.1.260 or later. On those surfaces:

• Run/advisorwith no argument to print the current advisor model and the aliases it accepts.
• Run/advisorwith a model, such as/advisor opus, to set it.
• Run/advisor offto turn it off.

Claude Code doesn’t invoke a saved advisor that your organization’savailableModelsallowlist excludes. To use the advisor, pick an allowed model with/advisor.Claude Code still saves an advisor that your current main model doesn’t support. That advisor activates after you switch to acompatible main modelwith/model. If the API already refused the saved advisor in the current conversation, it stays off until/clearor/compact, even after you switch models.On some plans, Fable as the advisor also needs your one-timeconsent to bill Fable usage to usage credits. For what/advisor fabledoes before you have given that consent, seeFable advisor and usage credits.

SetadvisorModelin settings

To configure the advisor as a default without opening a session, set it in your settings file:

Use the\-\-advisorflag

To set the advisor for a single session without changing your saved setting, launch with the flag:

Claude Code uses the flag instead of theadvisorModelsetting for that session. It doesn’t list\-\-advisorinclaude \-\-help. Claude Code exits with an error at launch if:

• The session’s main model doesn’t support the advisor
• The requested model, such as Haiku, can’t act as an advisor
• Your organization’savailableModelsallowlist excludes the requested model
• You requested Fable and your account still requires theusage-credits consent

If you start abackground sessionwith\-\-advisorand one of these applies, Claude Code starts the session without the advisor instead of exiting.If the requested model can act as an advisor butranks belowthe session’s main model, Claude Code starts the session anyway. Outside background sessions, it also warns at launch that the modelcannot advisethe main model.

Choose an advisor model

Claude Code ranks models by capability for the advisor role, and an advisor must rank at or above your session’s main model. Rows run from the lowest-ranked main model to the highest:

Fable 5.1 requires Claude Code v2.1.257 or later. Fable models requireFable access. Sonnet 5.5 as the advisor for an Opus 4.7 or Opus 4.8 main model requires Claude Code v2.1.287 or later.Set the advisor asfable,opus, orsonnet. These aliases resolve to Claude Code’sbuilt-in default versionfor each model family, which advances with new Claude Code releases. You can also pass a full model ID such asclaude\-opus\-5\-5. Haiku can call the advisor but can’t act as one.Subagents inherit the configured advisor and apply the same pairing check against their own model.Claude Code validates the pairing before sending a request, and the API validates it again:

• For an advisor that ranks below the main model, Claude Code doesn’t attach it to the main model’s requests. The/advisorcommand output and a notification show this; seeAdvisor is less capable than the current main model. Subagents whose own model satisfies the pairing may still use the advisor.
• If the API refuses the pairing of an advisor that Claude Code attached, Claude Code resends that request without the advisor. The conversation continues without one, so you see no error and get no advisor calls. If you then pick a different advisor with/advisor, it takes effect after/clearor/compactand in new sessions.
• If the main model or the advisor is a model Claude Code does not recognize, the advisor is not attached.

Fable advisor and usage credits

On some plans, Fable usage bills to usage credits, and Fable as the advisor bills the same way. If your account requires theone-time consent to bill Fable usage to usage credits, Claude Code asks for it when you select a Fable model with/modeland doesn’t apply Fable as the advisor until you have accepted that consent.Before you have accepted it, Claude Code doesn’t save Fable as the advisor when you type/advisor fableor pick Fable in the/advisorpicker. It points you to/model fableinstead. Withclaude \-\-advisor fable, Claude Code exits at launch with a message that points to/model fable. In abackground session, it starts the session without the advisor instead of exiting. With Fable already saved as youradvisorModel, Claude Code sends requests without the advisor. In an interactive session whose main model supports the advisor, it also shows a notification that points to/model fable.To accept the consent, run/model fableand choose to continue on Fable. Claude Code records the consent andsaves Fable as your selected model. Then select Fable as the advisor.

Common model pairings

Any accepted pairing works. These combinations balance cost against capability in different ways:

When Claude consults the advisor

Claude decides when to call the advisor. It tends to consult before committing to an approach, when an error keeps recurring, and before declaring a task done, but the timing is model-driven rather than rule-based.You can ask for a consultation in your prompt the same way you would request any tool, for exampleconsult the advisor before you continue. There is no setting to cap or force advisor calls; if you want Claude to consult more or less often during a task, say so in your instructions.

What you see during a session

When Claude calls the advisor, the transcript shows anAdvisingline with the advisor model name while the call is in progress. When the result returns, the line reports whether the advisor gave guidance:

• Reviewed: the line confirms that the advisor has reviewed the conversation. When the advisor returned readable guidance, pressCtrl\+Oto read it.
• Declined: the line readsAdvisor declined to advise on this request. If the advisor gave a reason, pressCtrl\+Oto read it.
• Unavailable: the advisor call failed, and the line readsAdvisor unavailable \(<error\_code\>\), where<error\_code\>is the code the call returned.

Claude generally follows the advisor’s guidance, but adapts when its own evidence contradicts a specific claim: if a recommended step fails when tried, or the file contents contradict the advice, Claude surfaces the conflict rather than following the guidance unconditionally.The advisor always receives the full conversation, and Claude controls the timing. For more control or a different configuration, seehow the advisor compares with subagents and opusplan.

Cost

When Claude calls the advisor, the advisor model reads the conversation, so each call consumes tokens at the advisor model’s rates in addition to your main model’s usage. How those advisor tokens are billed depends on how you pay:

• API billing: you pay the advisor model’s input and output rates for advisor tokens
• Subscription plans: advisor usage counts toward your plan’s usage limits, except that a Fable advisor bills tousage creditson plans where Fable usage does

If your account requires the usage-credits consent, a Fable advisor bills nothing before you give it, because Claude Codedoesn’t apply the selectionuntil then.Claude calls the advisor at decision points rather than on every turn, so pairing a faster main model with a stronger advisor typically costs less than running the stronger model throughout. Advisor usage counts toward the session totals shown by/usage.For how advisor tokens are reported in API responses, seeUsage and billingin the Claude API documentation.

Impact on prompt caching

Enabling or disabling the advisor mid-session does not invalidate your main model’sprompt cache. Unlikeswitching models, toggling/advisorkeeps the cached prefix intact, and the advisor’s returned guidance is cached as part of the transcript on later turns.The advisor model’s own read of the conversation is not cached. Each advisor call processes the full transcript anew, with no reuse between calls.

Requirements

The advisor tool requires all of the following:

• Anthropic API only: the advisor is a server-executed tool. It is not available on Amazon Bedrock, Claude Platform on AWS, Google Cloud’s Agent Platform, or Microsoft Foundry. Through anLLM gatewayconfigured withANTHROPIC\_BASE\_URL, availability depends on whether the gateway forwards the request intact to the Anthropic API. If the gateway or its upstream doesn’t recognize the advisor tool, seeAutomatic retry and error forwardingfor how Claude Code responds.
• Supported main model: Fable, Opus 4.6 or later, Sonnet 4.6 or later, or Haiku 4.5. SeeChoose an advisor modelfor which advisors each accepts.
• Feature-flag fetching: Claude Code turns the advisor on through a feature flag it fetches from Anthropic. In a session where a variable that turns flag fetching off is set, such asDISABLE\_TELEMETRY, the advisor stays off. SeeFeatures that need feature-flag fetching.

Turn the advisor off

To stop using the advisor, run/advisor offor chooseNo advisorin the/advisorpicker:

To disable the advisor tool entirely, setCLAUDE\_CODE\_DISABLE\_ADVISOR\_TOOL=1. The/advisorcommand becomes unavailable and any configuredadvisorModelis ignored. The\-\-advisorflag is accepted but has no effect. SeeEnvironment variables.

The advisor is one of several ways to combine model strengths. Pick based on when you want a second model involved.

See also

• Model configuration: switch models, set effort levels, and useopusplan
• Manage costs effectively: track token usage across models
• Advisor tool in the Claude API: understand the underlying server tool, or use it directly from the Messages API
• The advisor strategy: why pairing a fast main model with a stronger advisor works