The Agent Skills Directory

[INDIRECT_PROMPT_INJECTION]: The skill documents the ingestion of untrusted data from third-party social media and search platforms. It proactively identifies the risk of prompt injection from these sources and provides instructions to mitigate it, such as using explicit boundary tags like <FETCHER_UNTRUSTED_CONTENT> to isolate external data from agent instructions.
[EXTERNAL_DOWNLOADS]: The documentation references standard libraries for blockchain and HTTP interactions, including @x402/fetch, @x402/evm, @x402/svm, viem, and @solana/kit. These are legitimate dependencies for the payment and wallet integration functionality described in the skill.
[COMMAND_EXECUTION]: Includes standard curl examples for testing API endpoints and JavaScript fetch code for integrating the service. These are routine for API documentation and do not represent malicious command execution patterns.

fetcher