The Agent Skills Directory

[COMMAND_EXECUTION]: The skill executes local Node.js scripts and WebAssembly modules to perform API requests and cryptographic operations.
The file scripts/cloudveil-request.js is the primary entry point for executing CloudVeil API calls.
The function sm2Encrypt in scripts/cloudveil-request.js uses child_process.execFile to run node with a local bridge script and a .wasm file to perform SM2 encryption.
It reads sensitive credentials from environment variables (CLOUDVEIL_ACCOUNT, CLOUDVEIL_PASSWORD) and sends them (encrypted) to a user-defined CLOUDVEIL_BASE_URL.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through external data ingestion.
Ingestion points: Untrusted data enters the agent's context in scripts/cloudveil-request.js when fetching and parsing JSON responses from remote CloudVeil API endpoints.
Boundary markers: The skill lacks explicit boundary markers or instructions for the agent to ignore any natural language instructions that might be embedded in the API's JSON response fields (e.g., in a message or robotName field).
Capability inventory: The skill has significant capabilities, including issuing physical robot commands (delivery, transport, lifting, recharge), modifying system configurations (map switching, device shadow updates), and executing local shell commands via Node.js scripts.
Sanitization: There is no evidence of sanitization or validation of the content returned from the API before it is rendered or interpreted by the agent.

pudu-cloudveil-skill