Menü
Deep Research + Fragen-Katalog vor Implementierungsbeginn
Mit Codex oder Claude installieren Kopieren Sie diesen Prompt, fügen Sie ihn in Codex, Claude oder einen anderen Assistant ein und lassen Sie die Skill-Seite prüfen und installieren.
Multi-persona Round-Table mit domain-aware Personas (LLM wählt 4–6 themenspezifische Rollen); iterative Cross-Pollination bis TF-IDF-Konvergenz; Synthesizer ranked Top-N mit Pro/Contra
Research → Execute → Dual-Review Loop (Code Quality + Issue Resolution) bis beide Reviews bestanden
Multi-agent deep security audit — 6 expert personas (pentester, architect, code auditor, supply chain, data privacy, forensics) + CISO synthesis + optional fix implementation
Radical-honesty architectural review — questions code, methodology, design, and operational fitness
Security audit + fix — scans for hardcoded secrets, injection vectors, path traversal, input validation gaps, then implements all fixes
Cross-Domain Know-How-Transfer — findet kreative Anwendungen von Vault-Expertise in anderen Branchen
Basierend auf der SOC-Berufsklassifikation
| name | research-qa |
| description | Deep Research + Fragen-Katalog vor Implementierungsbeginn |
| version | 1.0 |
| requires | {"bins":[],"env":[],"os":[],"providers":[]} |
| tags | ["research","questions","planning","analysis","pre-implementation"] |
| config | {"timeout_minutes":60} |
Perform a structured pre-implementation research workflow:
Discovery: Explore the codebase thoroughly — read docs, directory structure, relevant source files, tests, configs, git history. Understand the project's architecture, patterns, and conventions. Do NOT modify any files.
Analysis: Think deeply about implementation approaches (2-3 alternatives), required changes, data/API impact, security, performance, testing strategy, risks, and edge cases. Do NOT modify any files.
Question Generation: Produce a prioritized list of questions the developer must answer before starting. Mark critical blockers with [BLOCKING]. Reference concrete code where relevant. Suggest options where possible.
Output written to .research-qa/. This is a read-only tool — no code changes.
