The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The repository's file structure contains cryptographic private key and secret key files in the tests/wallets directory, such as 0x20c1DF6f3310600c8396111EB5182af9213828Dc.eth.pk.txt and BTV1zY7njS5an91v9nphCK48d2vnMuecEgHLYiP25ycj.sol.sk.json. Storing actual credentials within the skill distribution represents a high-severity security risk.
[EXTERNAL_DOWNLOADS]: The documentation references multiple external blockchain services, including an Ethereum Sepolia gateway (https://sepolia.gateway.tenderly.co) that was flagged as a malicious URL by automated reputation scanners. Users should verify the integrity of this endpoint before use.
[INDIRECT_PROMPT_INJECTION]: The skill provides extensive capabilities for reading local files and uploading their content to decentralized storage, creating a vulnerability surface for indirect prompt injection.
Ingestion points: Data is ingested from the local file system through methods such as turbo.uploadFile and the upload-folder CLI command.
Boundary markers: No evidence of boundary markers or instructions to ignore embedded commands is provided in the documentation for processed content.
Capability inventory: The skill combines file system read permissions with network capabilities for data transfer and payment processing.
Sanitization: The documentation does not specify any sanitization or validation steps for the user-supplied content being processed.
[COMMAND_EXECUTION]: The CLI tools (turbo, npx turbo) take sensitive user arguments, including private keys and file paths, to trigger blockchain and file system operations. While these are core features, they require strict input handling to prevent command injection risks.

turbo-sdk