The Agent Skills Directory

[SAFE]: The skill is a standard developer tool for making HTTP requests in Go. It includes security-conscious features like the DefaultRedactor in its logging component to prevent the leakage of sensitive headers such as 'Authorization' and 'Cookie'.
[INDIRECT_PROMPT_INJECTION]: The skill inherently processes untrusted data from external HTTP responses (JSON, XML, etc.) and provides a capability to write content to the local filesystem via SaveToFile. While this presents an attack surface where remote data could attempt to influence the agent's behavior or persist malicious content, it is a standard function of an HTTP client library and does not include any specific exploitation patterns.
Ingestion points: HTTP response bodies processed in references/request.md (JSON, XML, Bytes) and references/body.md (Multipart, BodyGet).
Boundary markers: None identified for isolating external data from agent instructions.
Capability inventory: Network operations (Get, Post, Send in references/request.md) and file system writes (SaveToFile in references/request.md).
Sanitization: No specific sanitization or validation of remote content is described beyond standard decoding.

go-rc-fetch