The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill installs a binary from an external GitHub repository belonging to an unverified third-party author.
Evidence: The metadata and documentation specify the installation command go install github.com/Hyaxia/blogwatcher/cmd/blogwatcher@latest.
Context: Installing software from non-trusted repositories without specific version pinning poses a supply chain risk, as the source code could be modified to include malicious behavior.
[COMMAND_EXECUTION]: The skill requires the execution of a custom CLI tool to function.
Evidence: The skill body describes the use of blogwatcher commands such as add, scan, articles, and read.
Context: The security of the skill is dependent on the binary's integrity and its interactions with the host system.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to ingest and display content from external RSS and Atom feeds, which could serve as a vector for malicious instructions.
Ingestion points: External feed data retrieved and displayed via blogwatcher scan and blogwatcher articles in SKILL.md.
Boundary markers: No specific delimiters or instructions are provided to the agent to treat the external feed content as untrusted or to ignore embedded commands.
Capability inventory: The CLI tool performs network operations to fetch feeds and maintains a local record of blogs and articles.
Sanitization: The skill does not implement or describe any sanitization or validation of the data fetched from remote blogs before it is processed by the agent.

blogwatcher