santa-method

原文涵盖的主题

• Santa Method
• When to Activate
• Architecture
• Phase Details
• Phase 1: Make a List (Generate)
• Phase 2: Check It Twice (Independent Dual Review)
• Rubric Design
• Domain-Specific Rubric Extensions

Multi-agent adversarial verification framework. Make a list, check it twice. If it's naughty, fix it until it's nice. The core insight: a single agent reviewing its own output shares the same biases, knowledge gaps, and systematic errors that produced the output. Two independent reviewers with no shared context break this failure mode. Invoke this skill when: Output will be published, deployed, or consumed by end users Compliance, regulatory, or brand constraints must be enforced Code ships to production without human review Content accuracy matters (technical docs, educational material, customer-facing copy) Batch generation at scale where spot-checking misses systemic patterns Hallucination risk is elevated (claims, statistics, API references, legal language) Do NOT use for internal drafts, exploratory research, or tasks with deterministic verification (use build/test/lint pipelines for those). Execute the primary task. No changes to your normal generation workflow. Santa Method is a post-generation verification layer, not a generation strategy. Spawn two review agents in parallel. Critical invariants: 1. Context isolation — neither reviewer sees the other's assessment 2. Identic

这里只展示节选，请以完整原文和当前要求为准。