The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill fetches market data and yield pool information from trails-api.sequence.app. These network requests are used to resolve contract addresses for deposits and swaps and to verify real-time yield statistics.
[COMMAND_EXECUTION]: Instructs the agent to execute specific CLI commands for blockchain operations, such as agent swap, agent deposit, agent withdraw, and agent call. These commands are used to initiate transactions on various networks including Polygon, Ethereum, and Arbitrum.
[INDIRECT_PROMPT_INJECTION]: The skill relies on data from an external API which could influence the agent's decision-making process regarding transaction targets.
Ingestion points: Data is fetched from the Trails RPC API as described in the SKILL.md request examples.
Boundary markers: The skill does not explicitly define delimiters to separate the API's natural language fields from the agent's instructions.
Capability inventory: The skill uses the agent CLI to broadcast transactions to multiple blockchain networks, providing the agent with significant financial autonomy.
Sanitization: The skill defines a structured TypeScript interface to validate the incoming data's schema, though it does not specify additional filtering for the content of individual data fields.

polygon-defi