gabson0x/bountyforge/web3-grep-arsenal
Claude Code and agent skill by gabson0x/bountyforge
- community
Checking that the SKILL.md file is still in the repository.
Master grep command arsenal for Web3 smart contract auditing. Use when starting a new protocol scan, before deep code review, or when hunting specific vulnerability classes.
Install
npx vskill@latest install gabson0x/bountyforge/web3-grep-arsenalSKILL.md
GREP ARSENAL — MASTER REFERENCE
All grep commands in one place. Run in the first 30 minutes of any new target.
See also: [[Web3 Audit]], [[Web3 Bug Classes]], [[Methodology]], [[Trust Map]], [[Web3 Hunt Foundation]]
HOW TO USE THE SURFACE MAP
Process:
- Run ALL 10 blocks below (takes ~5 min)
- Collect all results in a notes file
- Tier-rank the hits (see Tier System below)
- In pass 1: READ everything, DON'T investigate yet
- In pass 2: Deep-dive on Tier 1 + 2 items
Tier System:
- Tier 1 — Near privileged code, external calls, or state changes with no guards → Investigate first
- Tier 2 — Interesting patterns that need context before judging → Investigate after Tier 1
- Tier 3 — Informational only (documentation, test files, comments) → Skip unless Tier 1+2 exhausted
THE 10 GREP BLOCKS (Copy-Paste Each)
Block 1 — Access Control
Red flags:
tx.originused for auth → Tier 1 (phishing vector)- Modifier uses
if (condition) { _; }without else → Tier 1 (silent bypass — function still executes for unauthorized callers) onlyOwnercount << total external function count → likely missing guards on siblings
Block 2 — Reentrancy
Red flags:
.call{value:}orsafeTransferBEFORE state updates in same function → Tier 1 (CEI violation)onERC721Received/onERC1155Receivedhooks present → check for reentrancy path- External calls present but
nonReentrantmissing → verify CEI is followed
Block 3 — Oracle / Price
Red flags:
slot0()used for price → Tier 1 (Uniswap V3 spot, flash-loan manipulable)getReserves()used for price → Tier 1 (Uniswap V2 spot, flash-loan manipulable)latestRoundDatawithoutupdatedAtcheck → Tier 1 (stale Chainlink price)latestAnswer→ Tier 1 (deprecated, no round validation)
Block 4 — Arithmetic / Math
Red flags:
unchecked {}blocks → manually verify each (Solidity 0.8+ unwraps here)- Division before multiplication (
a / b * c) → precision loss / 1e18in contract that handles 6-decimal tokens → decimal mismatch
Block 5 — Input Validation
Red flags:
delegatecallwith user-controlled target → Tier 1 (arbitrary code execution)abi.decodeon user-supplied calldata without length validation → Tier 1- Array params in batch functions without dedup check → Tier 1 (double-count attack)
Block 6 — Token Handling
Red flags:
token.transfer()withoutSafeERC20.safeTransfer()→ Tier 1 (return value unchecked, fails silently on old USDT)balanceOf(address(this))for pricing/shares → Tier 1 (donation attack vector)permit()without try/catch wrapper → Tier 2 (frontrun DoS possible)
Block 7 — ERC4626 / Vault
Red flags:
- ERC4626 present but
_decimalsOffset()NOT present → Tier 1 (first depositor inflation) totalAssets()usesbalanceOf(address(this))→ Tier 1 (donation attack)mint()ordeposit()called without same validation path → Tier 1 (MetaPool bug: mint skipped receipt check)
Block 8 — Proxy / Upgradeable
Red flags:
_authorizeUpgrade()withoutonlyOwneror role check → Tier 1 (anyone can upgrade)initialize()withoutinitializermodifier → Tier 1 (re-initialization possible)- Proxy present but
_disableInitializers()NOT in impl constructor → Tier 1 (impl attackable)
Block 9 — Signature / Replay
Red flags:
ecrecoverpresent butchainId/DOMAIN_SEPARATORNOT present → Tier 1 (cross-chain replay)ecrecoverwithout nonce → Tier 1 (same-chain replay)ecrecoverreturn not checked againstaddress(0)→ Tier 1 (invalid sigs succeed)
Block 10 — State Completeness / Access
Red flags:
- Role defined but
grantRole()call for that role NOT found anywhere → Tier 1 (role permanently empty) - Array-based function:
flag = trueOUTSIDE/AFTER loop → Tier 1 (empty array bypass) } catch { revert }on critical path → Tier 2 (liveness DoS if external changes)- Cache variable initialized to 0 with no first-access guard → Tier 1 (uninitialized cache)
PROTOCOL-SPECIFIC PATTERNS
Yield Aggregator (like Ern, Yearn, Beefy)
Lending Protocol (like Aave, Compound)
AMM / DEX
Staking / Restaking
ZK / Proof Contracts
2025 NEW PATTERN SCANS
SPECIFIC BUG PATTERN SEARCHES
Silent Modifier (if vs require)
This is the start of the file. Read the full SKILL.md on GitHub.
Security
- Security scanPassed, 100 of 100Pattern checks on the skill's files.
- Model reviewNot runA model review is scheduled for the next rescan.
Quality evaluation
- Task evaluationNot runNo evaluation yet. Run one in Skill Studio.
Source
- SKILL.md in the repositoryCheckingskills/web3/web3-grep-arsenal/SKILL.md
Popularity
- 7-day trend score
- 0
Not refreshed yet.
Works with
Any agent vskill installs to: 56 targets in all. These 39 have a page on this site.
- Claude Code
- Cursor
- GitHub Copilot
- Windsurf
- Codex
- Gemini CLI
- Cline
- Amp
- Roo Code
- Goose
- Continue
- Kilo Code
- Junie
- Trae
- OpenHands
- Qwen Code
- AI Agent
- Tabnine Chat
- Sourcegraph Cody
- Amazon Q
- Firebase Genkit
- Sweep
- Devin
- Phind
- Mentat
- Claude Desktop
- Aider
- GPT Pilot
- Plandex
- bolt.new
- v0
- Replit Agent
- Anse
- TypingMind
- BetterChatGPT
- LibreChat
- Warp AI
- Pieces
- Fig AI
Embed badge
[](https://verified-skill.com/skills/gabson0x/bountyforge/web3-grep-arsenal)