Loading
隐私声明： 我们使用 Cookie 来提升您的体验并分析网站流量。
基本 Cookie 是基本功能所必需的。分析 Cookie 帮助我们改进服务。
传统的智能合约审计仅关注代码安全性。虽然重要，但这种单一视角的方法忽略了导致大多数 DeFi 漏洞利用和项目失败的关键风险因素。
Spectra Audit 的multi-dimensional audit framework analyzes projects comprehensively across sentiment, code, tokenomics, liquidity, and distribution, providing a complete security picture that protects investors from threats code audits alone cannot detect.
每个维度提供独特的洞察。它们共同提供完整的安全智能。
对社交平台、论坛和开发者讨论中的社区情绪进行AI驱动的分析。
对智能合约代码进行深入的静态和动态分析，以识别漏洞和安全模式。
对代币分配、归属时间表、通胀机制和经济可持续性进行全面评估。
对DEX的流动性池、交易量、滑点容忍度和市场深度进行实时评估。
评估代币持有者集中度、巨鲸活动、空投分配和去中心化指标。
Every dimension is scored 0 to 100, multiplied by its published weight and divided by the weights we were able to use. The weights are fixed constants. They change when we change them and ship a new version, not while an audit is running.
The five dimensions below carry 75 of the 100 points available to a project with a token. The remaining 25 sit in the on-chain behaviour metrics listed in the full table further down.
Nothing tunes the weights per project and nothing adjusts them while an audit runs. They change only when we publish a new version of the scoring service, and the table on this page changes with it. Between versions we collect:
The worked example above assumes all ten dimensions produced a score. In practice some of them often cannot: a contract whose source is unverified yields no code score, a token with no holder history yields no whale dump risk score. Those dimensions are dropped, and the weights of the ones that remain are divided by their own sum so they still add up to 100. Nothing is scored zero for being unmeasurable.
These are the weights the scoring service applies, in full. Which column a project gets depends on whether it has a tradeable token: with a token, ten dimensions are weighted; without one, the six that need token data cannot be measured at all, so three remain.
| 维度 | 有代币 | 无代币 |
|---|---|---|
| 代码分析 | 35% | 55% |
| 代币分布 | 10% | 不适用 |
| 代币经济学 | 10% | 不适用 |
| 流动性 | 10% | 不适用 |
| 情绪 | 10% | 25% |
| Trading Velocity | 5% | 不适用 |
| Holder Growth | 5% | 不适用 |
| Realized PnL | 5% | 不适用 |
| Exchange Net Flow | 5% | 不适用 |
| Whale Dump Risk | 5% | 20% |
| Total | 100% | 100% |
A dimension we cannot measure for a given project is skipped rather than scored zero, and the weights of the dimensions that remain are divided by their own sum. A missing dimension therefore raises the share every other dimension carries; it never drags the score down by itself. If no dimension can be measured, no overall score is published.
我们的 AI 驱动系统按需跨所有评分维度分析项目。
实时摄取链上数据、社交信号、代码仓库和市场数据。
Advanced ML models process data across every scoring dimension the project supports.
Each dimension score is combined against the published weights, with the evidence behind every finding attached.
我们通过新维度和增强的研究能力扩展安全智能的旅程。
在 Q4 2025 之后，我们致力于根据新兴市场需求、技术进步和社区反馈，持续添加新的研究维度和增强分析因子。我们的框架旨在与区块链生态系统共同演进。
Experience the power of multi-dimensional analysis with weights you can read on this page. Get a comprehensive security audit in minutes, not weeks.
开始您的首次审计