The Agent Skills Directory

[COMMAND_EXECUTION]: The skill instructions guide the agent to construct and execute shell commands using the skale CLI tool. It uses templates for operations such as read, whitelist, and manager which are executed based on user-provided inputs.
[PROMPT_INJECTION]: The skill is vulnerable to indirect prompt injection as it interpolates external data into shell command strings. Ingestion points: User-provided arguments for contract methods, addresses, and parameters in examples/common-command-flows.md and references/command-matrix.md. Boundary markers: None identified; input is placed directly into the command structure without protective delimiters or instructions to ignore embedded commands. Capability inventory: The agent has the capability to execute shell commands via the CLI. Sanitization: Validation rules in rules/target-and-validation.md address Ethereum address formats and specific flags, but do not provide comprehensive filtering to prevent shell metacharacter injection or prompt overrides.

skale-cli