菜单
Apply systematic hypothesis validation when user formulates research claims or proposes testable hypotheses. Trigger on "I hypothesize that...", "Can you validate this claim?", or research findings requiring evidence assessment. Use confidence scoring and evidence tier classification.
用 Codex 或 Claude 帮你安装 复制这段 Prompt，粘贴到 Codex、Claude 或其他助手里，让它检查 Skill 页面并帮你完成安装。
基于 SOC 职业分类
| name | hypothesis-validator |
| description | Apply systematic hypothesis validation when user formulates research claims or proposes testable hypotheses. Trigger on "I hypothesize that...", "Can you validate this claim?", or research findings requiring evidence assessment. Use confidence scoring and evidence tier classification. |
| allowed-tools | Read, Grep, Glob, Edit, Write |
Validate research hypotheses using systematic methodology: assess testability, classify evidence, assign confidence levels.
Activate: "I hypothesize that...", "validate this claim", "what evidence supports", research claims, confidence assessment requests
Skip: Reading others' research (passive), well-established facts, brainstorming marked exploratory
| Tier | Type | Weight |
|---|---|---|
| 1 | Production deployments, measured outcomes | Highest |
| 2 | Peer-reviewed research, replication | High |
| 3 | Expert consensus, documented reasoning | Medium |
| 4 | Vendor claims, theoretical assertions | Low |
| 5 | Speculation, opinion, unverified | Minimal |
| Level | Criteria |
|---|---|
| HIGH (>80%) | 2+ Tier 1 sources OR multiple Tier 2 + Tier 1, no contradictions |
| MEDIUM (50-80%) | 1+ Tier 2-3 sources, some validation, minor contradictions |
| LOW (<50%) | Only Tier 4-5, theoretical only, significant contradictions |
# Hypothesis: [Title]
**Statement**: [Clear, testable hypothesis with scope]
**Confidence**: [HIGH/MEDIUM/LOW] ([XX]%)
**Evidence**:
- Tier 1: [Finding]
- Tier 2: [Finding]
**Contradictions**: [Any conflicts and resolution status]
**Language Guide**:
- HIGH: "demonstrates", "shows"
- MEDIUM: "suggests", "indicates"
- LOW: "hypothesizes", "may"
**Validation Plan**: [What's needed to increase confidence]
Monitors analysis-doc lifecycle in all three lanes — EMERGING→PRODUCTION promotion, PRODUCTION→RETIRING/RETIRED retirement (replacement-readiness as robust community/vendor elements mature), and follow-lane upkeep (followed canons' advance triggers + liveness, per ABSORPTION-MAP.md)
Validates new analysis documents against project quality standards. Trigger when user adds an analysis doc or says "review analysis".
Ensures SOURCES.md stays synchronized when adding new analysis documents or sources. Trigger when user adds a new analysis doc, says "update sources", "add source", or after pattern-reviewer identifies missing source entries. Enforces evidence tier classification and proper citation format.
Analyze recent git commits in a target project to surface patterns, identify improvement opportunities, and generate recommendations weighted by source authority and recency. Trigger when user says "analyze commits", "review my recent work", "what patterns do you see", or when evaluating a project for best-practice alignment.
Updates analysis documents with new version requirements and dependencies when Claude Code releases new features
Apply comprehensive quality review when user prepares content for publication (blog posts, articles, documentation, presentations). Trigger when user mentions "publish", "post", "article", "review this", "draft complete", "ready to share". Enforce evidence-based claims, intellectual honesty, professional voice, and balanced perspective.