security-smart-contracts

Audit Solidity and EVM smart contracts for vulnerabilities using static analysis, fuzzing, and manual review.

Updated Jul 9, 2026
One-click install
npx skills add https://github.com/akoita/agent-toolkit --skill security-smart-contracts-akoita
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: security-smart-contracts
Source: https://github.com/akoita/agent-toolkit/tree/main/plugins/claude/security/skills/security-smart-contracts
Command: npx skills add https://github.com/akoita/agent-toolkit --skill security-smart-contracts-akoita

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Smart contract code is immutable, publicly readable, and adversarially executed, so a single missed vulnerability can mean irreversible loss of funds. This Skill provides a structured four-phase audit procedure — reconnaissance, a two-pass sweep, deep validation, and reporting — that turns raw tool output into verified findings with written attack paths and proof-of-concept tests. ## Core Features & Use Cases - Four-phase audit workflow: Map roles and value flows, run syntactic and semantic sweeps, validate each candidate with a full call-chain trace, then emit a deduplicated severity-ranked report. - Toolchain orchestration: Copy-pasteable pipeline covering Slither, Aderyn, Solhint, Foundry invariants, Echidna, Medusa, Halmos, and hevm, with per-tool false-positive guidance and non-fatal exit handling. - 2026 vulnerability inventory: Reference material covering OWASP Smart Contract Top 10, reentrancy variants, oracle manipulation, upgradeability, ERC-4337, EIP-7702, Permit2, and agentic payment flows. - Use Case: Given a Foundry-based DeFi vault repository, run the default pipeline to generate recon and static-analysis artifacts, then validate each lead into a finding with a failing Foundry proof-of-concept test and a severity-ranked report. ## Quick Start Audit the Solidity contracts in this repository for security vulnerabilities and produce a severity-ranked findings report.

Frequently Asked Questions about security-smart-contracts

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I audit a Solidity smart contract for vulnerabilities?▼

Follow a four-phase procedure: map roles and value flows in reconnaissance, run syntactic and semantic sweeps, validate each candidate by tracing its full call chain, then report confirmed findings with severity and proof of concept. Static analyzers like Slither and Aderyn generate leads, but every lead needs a written attack path before promotion.

Slither vs Aderyn for smart contract static analysis?▼

Run both, because they ship different detector sets and their disagreement is the interesting part. Slither offers richer printers and companion binaries for upgradeability and storage checks, while Aderyn produces lower false-positive volume and an MCP server. Deduplicate results by file, line, and vulnerability class.

How do I write Foundry invariant tests for a DeFi protocol?▼

Target a handler contract that bounds inputs with bound() rather than vm.assume, tracks ghost variables, and rotates through an actor array. Assert the six invariant families: conservation, solvency, monotonicity, access control, round trip, and no stuck funds. Set fail_on_revert to true and configure max_time_delay for time-dependent logic.

Does Solidity 0.8 checked arithmetic prevent all overflow bugs?▼

No. Checked arithmetic covers ordinary expressions only. Unchecked blocks, inline assembly, and type downcasts like uint128(x) still wrap or truncate silently at every compiler version. Every downcast and unchecked block remains an audit candidate regardless of pragma version.

When should I not use smart contract audit tooling?▼

Skip this workflow for off-chain application, cloud, container, or dependency security, which belong to other review processes. It also does not cover gas optimization, tokenomics or economic-design review without a security question, deep non-EVM chain analysis, or contracts you are not authorized to review.