etherscam
v1.0.9
Published
AI-powered smart contract vulnerability analyzer using multi-model detection
Maintainers
Readme
etherscam
AI-powered smart contract vulnerability analyzer using multi-model detection.
Features
- Multi-model analysis pipeline (Gemini, Mimo, Devstral, Grok)
- Web research for external protocol dependencies (Uniswap, Aave, Chainlink, etc.)
- MEV threat modeling (frontrunning, sandwich attacks, multi-block MEV)
- Economic and game-theory vulnerability detection
- Post-deployment remediation strategies
- Secure API key storage via system keychain
Requirements
- Bun >= 1.0.0
- 4everland API key (get one here)
Installation
bun install -g etherscamOr clone and link locally:
git clone https://github.com/angrymouse/etherscam
cd etherscam
bun install
bun linkSetup
Configure your API key (stored securely in system keychain):
etherscam --setupUsage
etherscam <contracts-folder>Options:
-v, --verbose- Show detailed analysis including codebase overview-s, --setup- Reconfigure API key-h, --help- Show help
Example:
etherscam ./contracts -vHow It Works
- Phase 1: Flow Extraction - Identifies contract entrypoints and user interaction flows
- Phase 2: Codebase Overview - Generates technical overview via model chain (devstral → mimo → gemini)
- Phase 3: Multi-Model Detection - Four models analyze each flow for vulnerabilities with web research capability
- Phase 4: Report Generation - Gemini Pro synthesizes findings into a structured security report
Output
Generates REPORT.md in the target folder containing:
- Executive summary
- Confirmed vulnerabilities with file:line locations
- Economic and logic issues
- Other potential issues (possible false positives)
- Recommendations
- Post-deployment remediation (admin actions, whitehat rescue, guardian bots, migration plans)
License
MIT
