claude

Model Crosscheck

Multi-model workflows for Claude Code: /dual-plan and /tri-plan (rival models critique a plan until convergence), /dual-review and /tri-review (independent reviewers on the same diff, merged and adversarially verified), /tri-decide (blind proposals for hard-to-reverse architecture decisions), /tri-strategy (three assigned lenses over one evidence pack for business decisions), and /tri-research (one researcher, two auditors, and a claim ledger they verify row by row).

10 on GitHubby Sameer Ahmed Khanv2.14.0MITupdated 6 days ago
/

Plugin crosscheck from marketplace model-crosscheck in SameerKhan/model-crosscheck. Commands read from the repo's manifest — 5 days ago.

What it installs

Skills (7)

  • dual-plan

    Plan and build a feature with Claude AND OpenAI Codex together: Claude drafts the implementation plan, Codex critiques it against the codebase, Claude merges and implements after user approval, then /dual-review gates the merge. Use when the user says "dual plan", "plan with both", "plan this with C

  • dual-review

    Dual code review: run Claude's /code-review AND OpenAI Codex CLI review on the same diff, then merge both into one consolidated report. Use when the user says "dual review", "review with both", "Claude + ChatGPT review", or before a merge when they want two independent reviewers.

  • tri-decide

    Decide an architecture or technology question with Claude, OpenAI Codex, and Google Gemini proposing independently and blind, then cross-examining each other's approaches anonymized. Produces a decision record: options, cruxes, what each bets on, a recommendation, and a revisit trigger, not a winner

  • tri-plan

    Plan a feature with Claude, OpenAI Codex, AND Google Gemini together: Claude drafts the implementation plan, Codex and Gemini independently critique it against the codebase in parallel, Claude adjudicates and implements after user approval, then /tri-review gates the merge. Use when the user says "t

  • tri-research

    Research a factual question with Claude, OpenAI Codex, and Google Gemini: Claude gathers the facts and writes a claim ledger (every number dated and sourced), the other two re-fetch the sources and try to break it, then Claude re-verifies every disputed row against the page itself. Produces one ledg

  • tri-review

    Triple code review: run Claude's /code-review, OpenAI Codex CLI, AND Google Gemini (Antigravity CLI) on the same diff, then merge all three into one consolidated report. Use when the user says "tri review", "triple review", "review with all three", "Claude + Codex + Gemini review", or before a merge

  • tri-strategy

    Decide a business or strategy question with Claude, OpenAI Codex, and Google Gemini arguing from three deliberately different lenses: unit economics, competitive positioning, and execution capacity, over the same evidence pack. Produces a strategy memo with what each option bets on, the cruxes resol

View source on GitHub

Maintain Model Crosscheck?

This page updates itself from your repo. Add the badge to your README, or get it Featured to pin it above the organic results with a followed link to your site.

claude & codex pluginslisted