The Agent Skills Directory

[CREDENTIALS_UNSAFE]: The skill requires users to manage an OPEN_TOKEN and provides functionality to transmit sensitive exchange API keys (apiKey, secret, password) to the vendor's backend for configuration. While the documentation advises against committing these to version control, the agent is actively involved in handling and transmitting raw credentials.
[PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through its data ingestion surfaces.
Ingestion points: Market ticker data (/market/ticker), K-line data (/market/klines), and proposed workflows (Workflow C and F) that ingest external news and social media content.
Boundary markers: No explicit delimiters or instructions are provided to the agent to ignore potential malicious commands embedded in external market news or sentiment data.
Capability inventory: The skill possesses the ability to execute network operations (curl) and perform financial transactions (POST /orders).
Sanitization: There is no evidence of sanitization or validation of the content received from third-party news or social media sources before it influences trading decisions.
[COMMAND_EXECUTION]: The skill documentation and operation flow rely on executing curl commands to interact with the API endpoints for market data, account management, and trade execution.
[EXTERNAL_DOWNLOADS]: The skill makes network requests to ai.6551.io to perform all functions, including configuration updates and trading operations. This is a vendor-controlled domain associated with the skill author.

opentrade-cex