The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and process data from a wide range of external blockchain and market data providers, including Alchemy, Moralis, CoinGecko, and Hyperliquid. This processing of untrusted external content represents a potential surface for indirect prompt injection attacks.
Ingestion points: Data enters through the invoke_api_endpoint and invoke_stateful_endpoint tools (SKILL.md, references/root-mcp-workflow.md).
Boundary markers: The skill requires an 'Evidence receipt' and uses the validate_task_result tool to verify the structure of server responses, which serves as a defensive boundary (SKILL.md).
Capability inventory: The skill possesses the capability to perform both read operations and stateful changes (writes) on blockchain networks via MCP tools.
Sanitization: The instructions mandate structured reporting and structural validation via validate_task_result, though it notes this does not authenticate the content itself.

hive-query