MCP

Tools

The get_verdict tool schema and output format.

get_verdict

Run a hard decision through the solvemax engine: clarify, widen options, locked-weight scoring, 5,000 simulated futures, red team, blind audit. Returns the full verdict slip (headline, podium, dual verdict, kill criteria, verification) plus a shareable URL. Costs 1 solvemax credit.

Parameters

ParameterTypeRequiredNotes
decisionstringyes20–6,000 characters. The decision in plain words — options optional; include numbers, constraints, and feelings.
mode"DECIDE" | "COMPARE" | "GRILL"noGRILL = attack a plan; COMPARE = rank supplied options; default is auto-detect. See Decision runs.

Every call sets auto: true on the underlying API request — the engine answers its own clarifying questions with its recommended picks, so the tool returns a complete verdict in a single call with no follow-up round trip for the AI assistant to manage.

Output

Plain text, formatted for inline chat display:

VERDICT: <headline>

<summary>

Win rate: <leaderWinRate>% of 5,000 simulated futures (ROBUST|CLOSE CALL) · confidence <confidence>%
Podium: 1. <name> (<winRate>%)  2. <name> (<winRate>%)  3. <name> (<winRate>%)
Verification: <verdict> — independent winner: <independentWinner>
Exit plan: <killCriteria joined by " | ">
Trade-off: <tradeoff>

Full slip: <url>   credits left: <credits_remaining>

Errors

If SOLVEMAX_API_KEY isn't configured, or the underlying API call fails, the tool returns an error result with the same error code and detail the API would return — nothing is swallowed silently.

Cost

Each call spends 1 credit from the account the API key belongs to — the same credit balance used everywhere else. See Credits.

On this page