The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill metadata and documentation facilitate the installation of software from a third-party GitHub repository (github.com/Hyaxia/blogwatcher). This source does not originate from the documented vendor infrastructure.
[REMOTE_CODE_EXECUTION]: The skill facilitates remote code execution through the go install command, which downloads, compiles, and installs code from a remote repository. The skill relies on this externally sourced binary to function.
[COMMAND_EXECUTION]: The skill's primary functionality is derived from executing the blogwatcher CLI tool. This involves spawning subprocesses and interacting with the local operating system to manage data and fetch network resources.
[PROMPT_INJECTION]: The skill exhibits an attack surface for indirect prompt injection by processing untrusted content from external RSS and Atom feeds.
Ingestion points: Untrusted data enters the agent context via the blogwatcher scan and blogwatcher articles commands, which fetch content from arbitrary URLs.
Boundary markers: The skill does not provide delimiters or instructions to the agent to ignore or sanitize embedded instructions within the feed data.
Capability inventory: The skill possesses the ability to execute CLI commands and perform network operations.
Sanitization: There is no evidence of content sanitization, filtering, or validation before the feed data is processed by the agent.

blogwatcher