The Agent Skills Directory

[COMMAND_EXECUTION]: The script scripts/radius-wallet-bootstrap.mjs executes the Foundry cast command to generate local development wallets. This is a legitimate development utility that includes security controls such as regex-based name validation, restricted file permissions (0600), and automated .gitignore management to prevent accidental disclosure of secrets.\n- [EXTERNAL_DOWNLOADS]: The skill references official Radius documentation and RPC endpoints at radiustech.xyz. These references are used for configuration and reference data and do not involve untrusted code execution or third-party dependencies.\n- [CREDENTIALS_UNSAFE]: The skill enforces best-practice guidance for secret management, specifically advising against leaking private keys in shell history or process listings. It provides a bootstrap script that secures generated keys in local, git-ignored files with restricted permissions, correctly separating sensitive key material from application logic.\n- [PROMPT_INJECTION]: The skill identifies a potential surface for indirect prompt injection via the ingestion of external documentation URLs.\n
Ingestion points: Documentation files at docs.radiustech.xyz linked in SKILL.md.\n
Boundary markers: SKILL.md includes an explicit 'Trust boundary' warning that instructs the agent to treat fetched content as reference data only and to ignore any instructions or tool calls embedded within it.\n
Capability inventory: The skill utilizes command execution (cast via the bootstrap script), file system access for wallet management, and network RPC interaction.\n
Sanitization: The bootstrap script performs regex-based sanitization on user-supplied wallet names to prevent command injection.

radius-dev