메뉴
Multi-perspective reasoning (advocate/critic/synthesizer). Three modes - general, --code for architecture decisions, --medical for EBM clinical analysis.
Codex 또는 Claude로 설치 이 Prompt를 복사해 Codex, Claude 또는 다른 어시스턴트에 붙여 넣으면 Skill 페이지를 검토하고 설치를 진행할 수 있습니다.
SOC 직업 분류 기준
Code architecture design with web-informed debate between Architect and Reviewer agents — produces ASCII diagrams, justified stack, patterns (SOLID/DDD/CQRS), phased implementation plan.
Multi-turn brainstorming session with RAG, domain profiles (code/tech/btp/cyber/finance), Claude Opus or local Qwen3.5 backend.
CSV analysis via 2-pass SOTA orchestrator (score 54/55 on 7 CSVs). Statistics, data profiling, visualizations, insights.
Deep reflection with web search — Brave/Perplexica pipeline fed into Qwen3.5-35B thinking mode for grounded, well-reasoned answers.
Threat intelligence analysis of IoCs (IPs, domains, hashes, URLs) via CyberBro (33 engines including Shodan/AbuseIPDB/VirusTotal) + Brave Search enrichment.
MCP CLI — list, configure, auth, and call MCP servers/tools directly (HTTP or stdio), including ad-hoc servers, config edits, and CLI/type generation.
| name | debate |
| description | Multi-perspective reasoning (advocate/critic/synthesizer). Three modes - general, --code for architecture decisions, --medical for EBM clinical analysis. |
| trigger_patterns | ["débat","pour ou contre","faut-il","vs","arguments","trade-offs"] |
| tools_required | ["exec","llm_qwen35"] |
| examples | ["Faut-il courir un marathon sans entrainement ?","--code Should we use PostgreSQL or MongoDB for user session storage?","--code REST vs GraphQL for a mobile-first API with 50+ clients?","--medical Le TENS est-il indiqué pour la lombalgie chronique ?","--medical Faut-il prescrire de la vitamine D systématiquement en hiver ?"] |
| model | qwen3.5-35b-a3b |
| execution | {"command":"bash /home/remondiere/.openclaw/bin/debate_wrapper.sh '$args'","timeout_ms":300000,"arg_mode":"raw"} |
Débat structuré en 3 passes séquentielles sur Qwen3.5-35B. Trois modes disponibles.
Trois passes en français :
Trois passes en anglais, optimisées pour la prise de décision technique :
Trois passes en français, optimisées pour l'analyse clinique basée sur les preuves :
debate_router.py [--code|--medical] [--verbose] [--json] [--context TEXT] <question>
--verbose / -v : affiche les 3 passes complètes + timings--json : output JSON {mode, debate: {advocate, critic, synthesis}, elapsed_s}--context / -c : injecte un contexte RAG dans chaque passe| Mode | Pass 1 | Pass 2 | Pass 3 | Total |
|---|---|---|---|---|
| general | ~15s (no-think) | ~30s (think) | ~40s (think) | 60-90s |
| --code | ~15s (no-think) | ~30s (think) | ~45s (think) | 70-100s |
| --medical | ~15s (no-think) | ~30s (think) | ~40s (think) | 60-90s |
Timeout: 300s (120s par passe). Passes 2 et 3 toujours en thinking mode.