ngmeyer

Run any question, plan, PR, or code through a Diverse Multi-Agent Debate (DMAD) council of 5 AI advisors with distinct reasoning methods. Advisors collaborate, peer-review each other anonymously, and a chairman synthesizes a verdict, empirically outperforming adversarial debate.

Performs multi-source deep research, generating cited, confidence-scored briefs by searching Exa, Tavily, and the web in parallel. It extracts claims, scores confidence, and highlights contradictions, working without API keys and improving with each added provider.

Reconcile session outcomes into persistent project memory files, updating project state, backlog, and status via section-aware merging. Use this at the end of any significant work session, especially when closing or saving a session.

The "audit" mode identifies drift (claimed facts no longer matching code), leaked secrets, duplicates, instruction-budget bloat, and prescriptive-vs-descriptive imbalance across all CLAUDE.md files in your projects.

Generates decision memos and product launch documents in Amazon's narrative style. It offers two modes: 'memo' for 6-page narrative memos and 'prfaq' for press releases and FAQs, enforcing Strunk's prose rules.

Forge new Claude Code skills the right way, incorporating proper frontmatter, progressive disclosure, helper scripts for deterministic work, mandatory Gotchas, and an iterate-then-extract meta-process. Use this when creating, writing, building, scaffolding, or improving a skill.