web3-smart-contracts

Pass

Audited by Gen Agent Trust Hub on Mar 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [SAFE]: The skill is primarily focused on educational and auditing tasks for Web3 development. It strongly promotes security-first principles, such as the Checks-Effects-Interactions (CEI) pattern and the use of battle-tested libraries like OpenZeppelin. No evidence of malicious intent, data exfiltration, or obfuscation was found.
  • [COMMAND_EXECUTION]: The 'Companion check' feature in SKILL.md instructs the agent to execute a directory listing command: ls ~/.claude/skills/ ~/.agent/skills/ ~/.agents/skills/ .claude/skills/ .agent/skills/ .agents/skills/ 2>/dev/null. This is a read-only discovery operation used to identify which of the author's other skills are installed so it can recommend missing ones. It does not access sensitive files or credentials.
  • [EXTERNAL_DOWNLOADS]: The skill suggests the installation of companion skills using npx skills add AbsolutelySkilled/AbsolutelySkilled --skill <name>. This is a standard installation pattern using a well-known package runner (npx) and points to the vendor's official resources.
  • [PROMPT_INJECTION]: The skill analyzes untrusted user-provided smart contract code, creating a surface for indirect prompt injection.
  • Ingestion points: Solidity source code and contract files provided by the user for review or auditing.
  • Boundary markers: Relies on the host agent's native context separation.
  • Capability inventory: Shell access for benign environment discovery (ls) and suggested tool installation (npx).
  • Sanitization: None explicitly mentioned in the skill instructions.
  • Analysis: This surface is inherent to the skill's purpose as a code auditor. The skill does not grant the agent capabilities to perform dangerous operations or exfiltrate data based on the processed code.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 23, 2026, 01:04 PM
Security Audit — agent-trust-hub — web3-smart-contracts