The Agent Skills Directory

[COMMAND_EXECUTION]: The skill extensively uses the mplx CLI to perform blockchain operations, including asset creation, transfers, and wallet management on the Solana network.
[DYNAMIC_EXECUTION]: Provides functionality to execute arbitrary Base64-encoded blockchain instructions via the mplx toolbox transaction command and equivalent SDK methods, allowing the agent to perform any on-chain action supported by the underlying protocols.
[INDIRECT_PROMPT_INJECTION]: The skill exhibits an indirect prompt injection attack surface by processing data from untrusted external sources.
Ingestion points: Processes NFT and token metadata in JSON format from local files and remote URIs (Arweave/Irys) as documented in metadata-json.md, cli-core.md, and sdk-umi.md.
Boundary markers: Absent; there are no instructions or delimiters defined to prevent the agent from obeying instructions embedded within the metadata fields (e.g., name, description).
Capability inventory: High-privilege capabilities including the ability to transfer funds, burn assets, modify collection settings, and execute raw transactions.
Sanitization: Absent; the skill does not specify sanitization or validation logic for data ingested from external URIs before it is used to influence agent decisions.
[EXTERNAL_DOWNLOADS]: Fetches the Metaplex CLI and various official SDK packages from established registries like NPM. These downloads originate from the official Metaplex Foundation and Solana organizations.

metaplex