The Agent Skills Directory

[COMMAND_EXECUTION]: Provides comprehensive documentation for interacting with the http and https command-line interfaces.
[DATA_EXFILTRATION]: Outlines the tool's capability to read local file content and incorporate it into network requests using separators like @, =@, :=@, :@, and ==@ as documented in references/request-items.md.
[CREDENTIALS_UNSAFE]: Describes how the tool manages sensitive data in local configuration and session storage, such as ~/.netrc and persistent session JSON files.
[REMOTE_CODE_EXECUTION]: Details the installation of third-party plugins using the httpie cli plugins install command, which allows extending tool functionality via external Python modules.
[INDIRECT_PROMPT_INJECTION]: Identifies the attack surface where the agent processes data from external URLs.
Ingestion points: HTTP response bodies received from arbitrary URLs (documented in references/cli.md and references/scripting.md).
Boundary markers: None mentioned.
Capability inventory: Shell command execution, file writing, and plugin installation (documented across references/cli.md and references/plugins.md).
Sanitization: Not addressed in the reference documentation.

httpie