The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to analyze external, potentially untrusted smart contract source code (Solidity and Rust). This creates a vulnerability surface where malicious instructions could be embedded in the code being audited to influence the agent's behavior. * Ingestion points: The skill performs analysis on files within the src/ directory (*.sol, *.rs). * Boundary markers: The skill includes a 'Mandatory confirmation gate' that requires the agent to verify targets, authorization, and commands with the user before proceeding. * Capability inventory: Includes shell commands (grep) and blockchain CLI tools (solana, spl-token) to inspect contract state and source code. * Sanitization: The skill relies on static pattern matching and user confirmation rather than runtime execution of the audited code.
[COMMAND_EXECUTION]: The skill provides various shell commands for finding vulnerability patterns (e.g., searching for minting functions, blacklist logic, or fee manipulation). These commands are intended to be executed in the user's environment to assist in the audit process. The skill explicitly instructs the agent to wait for user confirmation before running these commands.

meme-coin-audit