Entry Point Analyzer maps the attack surface of a smart contract project. It detects the language (Solidity, Vyper, Solana, Move for Sui and Aptos, TON, CosmWasm), lists externally callable functions that change state, and classifies each one as public, role-restricted, restricted pending review, or contract-only.
Read-only functions are left out on purpose. The result is a structured markdown report with file and line references, ready to guide an audit.
What You Get
- A markdown report with counts per access category
- Tables of public, role-restricted, review-required and contract-only entry points with file and line
- Notes on modifiers and inherited access control for each function
- A list of files analyzed and any that could not be parsed
Common Use Cases
- Starting a smart contract audit by mapping what anyone can call
- Listing privileged operations such as upgrades, pausing and fee setters
- Checking that callbacks really restrict their caller
- Scoping a review to one directory of a large protocol
What Makes It Unique
It covers six contract ecosystems in one skill, with seven reference files, and it classifies access with care: a callback name or an msg.sender != tx.origin check does not count as access control, and unclear cases are flagged for review so the auditor makes the call. It also lists rationalizations to reject, such as “the modifier name is clear”.
Under The Hood
The skill is prose plus per-language reference files and no bundled scripts. The agent reads and searches the contract source and, for Solidity, checks whether Slither is installed with which and runs its entry-points printer in the project folder. If Slither is missing or fails, it falls back to manual reading. It does not fetch web pages and writes only the report it produces.