The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill uses npx to fetch and execute the @morpho-org/cli package from the official NPM registry. This package is provided by the protocol vendor for querying data and preparing transaction payloads.
[COMMAND_EXECUTION]: All protocol interactions are performed via shell commands. The instructions provide structured schemas for querying vault/market data and preparing transaction objects.
[SAFEGUARD_OBSERVED]: The skill implements several security boundaries: it explicitly prohibits the agent from handling private keys or signing transactions, mandates checking simulation status (simulationOk) before presenting results to users, and requires reporting CLI errors directly to the user without attempting workarounds.
[PROMPT_INJECTION]:
Ingestion points: The agent processes data retrieved from the Morpho protocol (such as vault names and market descriptions) via CLI JSON output.
Boundary markers: No specific delimiters are used to separate external tool output from the internal prompt context.
Capability inventory: The agent has the ability to execute shell commands to prepare transactions based on protocol data.
Sanitization: The skill relies on the structured nature of the CLI's JSON output and does not specify explicit sanitization for string fields returned from the blockchain.

morpho-cli