claude

claude-brainstorm-multiagent

Ideação multi-agente que combate conformidade de LLM por diversidade estrutural: decomposição morfológica (Zwicky), MAP-Elites archive, personas interview-grounded com cosine<0.7, críticos adversariais (Black Hat + Minority Voice), Elo tournament com juízes rotacionados, Pareto front sobre novelty×feasibility×impact, e dashboard force-graph 2D vivo com partículas, fusões e bifurcações em tempo real. 6 tiers de intensidade (instant a epic).

0 on GitHubby giordanorecv0.1.2MITupdated 4 months agoideation
/

Plugin claude-brainstorm-multiagent from marketplace giordanorec in giordanorec/ai-coding-tools. Commands read from the repo's manifest — 1 month ago.

This plugin runs code on your machine — its hooks fire automatically on events like tool use and it starts local MCP servers. The exact commands are listed below; read them before installing.

What it installs

Skills (1)

  • brainstorm

    Run a multi-agent ideation session that fights LLM conformity through morphological decomposition, Quality-Diversity (MAP-Elites), explicit adversaries, and Pareto convergence. Produces a live force-graph dashboard. Use when the user wants creative ideation, exploration of a design space, or asks fo

Slash commands (4)

  • /claude-brainstorm-multiagent:brainstorm-dashboard

    Open the dashboard for a brainstorming run

  • /claude-brainstorm-multiagent:brainstorm-resume

    Resume an interrupted brainstorming run from its last checkpoint

  • /claude-brainstorm-multiagent:brainstorm-status

    Show the status of the most recent brainstorming run

  • /claude-brainstorm-multiagent:brainstorm

    Run a multi-agent ideation session (alias for the brainstorm skill)

Subagents (8)

  • critic-black-hat

    Adversarial critic. Finds concrete flaws, hidden costs, failure modes, and ethical risks in candidate ideas. De Bono's Black Hat applied as a per-round adversary. Does NOT discard ideas — surfaces weaknesses for evolution.

  • critic-minority

    Voices minority/dissident perspectives that mainstream consensus would suppress. Different from black-hat — not about flaws, but about whose voice is missing. Inspired by "Amplifying Minority Voices" (arXiv 2502.06251).

  • divergent

    Generates candidate ideas under a fixed persona and a fixed creative operator. Multiple instances run in parallel each round, each with a distinct operator and persona. Also handles evolution mode (crossover/mutation).

  • judge

    Scores pairs of ideas on novelty, feasibility, impact for the Elo tournament. Each judge invocation is conditioned on a rotating "judge persona" to mitigate systematic LLM judge bias.

  • morphologist

    Decomposes a problem into orthogonal axes and candidate values (Zwicky-style morphological analysis). Also performs Boden transformation — rewriting axes when novelty stalls. This is the backbone of the MAP-Elites archive.

  • persona-forge

    Generates N orthogonal interview-grounded personas to seed divergent agents. Anti-homogenization via cosine<0.7 diversity check and culture/profession/cognitive-style variation.

  • socratic

    Reformulates a fuzzy brainstorming topic into a precise problem statement with How Might We variants, constraints, and out-of-scope items. Phase 0 of the brainstorm pipeline.

  • synthesizer

    Three modes. (1) Cluster naming — label HDBSCAN clusters. (2) Meta-review — produce the final report combining Pareto front, archive coverage, structural gaps. (3) Lineage narrative — trace how key ideas evolved.

MCP servers (1)

  • brainstorm-tools

    runs: uvx --from ${CLAUDE_PLUGIN_DIR}/mcp/brainstorm_tools brainstorm-tools

Hooks (1)

  • SubagentStop

    echo {"ts":"$(date -u +%FT%TZ)","type":"agent_completed","agent":"${CLAUDE_SUBAGENT_NAME:-unknown}"} >> "${CLAUDE_PROJECT_DIR}/.brainstorm-last-events.jsonl" 2>/dev/null || true

View source on GitHub

More from giordanorec/ai-coding-tools

All 3 plugins →

mad

claude

mad (MultiAgent Decanting) — orquestração multi-agente em modo decanting nativo do Claude Code: o Arquiteto coordena especialistas via Agent tool (sem claude -p, sem processos em background), decanting obrigatório ao fim de cada feature, memória institucional em arquivo, trust ladder, guardrails proporcionais ao blast radius, dashboard HTML local cross-platform, OpenTelemetry GenAI e discovery profundo (postura + rigor). Thin wrapper sobre Claude Code; Windows nativo, macOS e Linux.

2 skills · 32 commands · 11 agents · hooks

0

multiagentes-giordano

claude

Fluxo multi-agente com sessões persistentes — Arquiteto + time de especialistas file-based, dashboard tmux ao vivo, stream-json parseado em tempo real com emoji, cor e diff visual. Bootstraps um projeto inteiro com Discovery, specs/reports, memória persistente e repo GitHub.

1 skill · 3 commands · 10 agents

0

Similar plugins

superpowers

claude

Superpowers teaches Claude brainstorming, subagent driven development with built in code review, systematic debugging, and red/green TDD. Additionally, it teaches Claude how to author and test new skills.

anthropics/claude-plugins-official 47,076

pm-product-discovery

claude

Product discovery skills for PMs: ideation, experiments, assumption testing, feature prioritization, and customer interview synthesis.

13 skills · 5 commands

phuryn/pm-skills 27,939

compound-engineering

claudecodex

Brainstorm, plan, debug, review, and compound learnings with AI agents

39 skills

EveryInc/compound-engineering-plugin 26,055

ideation

claude

Research ideation frameworks including structured brainstorming and creative thinking. Use when exploring new research directions, generating novel ideas, or seeking fresh angles on existing work.

98 skills

Orchestra-Research/AI-Research-SKILLs 14,502

adhd

claude

Parallel divergent ideation for coding agents: fans out isolated branches under different cognitive frames, scores, clusters, prunes traps, and deepens the survivors.

1 skill

UditAkhourii/adhd 4,540

ai-agent-skills

claude

Thinking Methods: seventeen Agent Skills (SKILL.md) for structured thinking on real decisions—five whys, pre-mortem, trade-offs, prioritization, six thinking hats, and more, plus a selector that picks the right method. Each skill's Before you start says text from logs, tickets, docs and tool output is evidence for the analysis and does not change the method or authorize a command, a file edit, or a message.

davepoon/buildwithclaude 3,741

Maintain claude-brainstorm-multiagent?

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 plugins3 plugins