The Agent Skills Directory

[DATA_EXFILTRATION]: The skill provides instructions for sending data to external storage endpoints, such as publisher.walrus-testnet.walrus.space. While this is the intended purpose of the storage skill, it establishes a mechanism for transmitting potentially sensitive data to an external service.- [EXTERNAL_DOWNLOADS]: The documentation references and recommends the use of the @mysten/walrus Node.js package for mainnet operations.- [COMMAND_EXECUTION]: The skill utilizes curl commands to interact with remote HTTP APIs for reading and writing data blobs.- [PROMPT_INJECTION]: The skill is designed to fetch and process data from a decentralized network. Because the content of these blobs is untrusted, it could contain malicious instructions designed to manipulate the agent's behavior (Indirect Prompt Injection).
Ingestion points: Data fetched via curl from Walrus aggregator endpoints (e.g., aggregator.walrus-mainnet.walrus.space).
Boundary markers: No specific boundary markers or instructions to ignore embedded commands are provided for the ingested blob content.
Capability inventory: The skill uses curl for network requests; the agent may have further capabilities to write or process the resulting data.
Sanitization: The instructions do not include steps for sanitizing or validating the content received from the storage network before processing.

walrus