The Agent Skills Directory

[SAFE]: The skill provides a purely instructional framework for the AI agent to follow a specific reasoning process. It does not request or use any tools that interact with the operating system, file system, or network.
[NO_CODE]: The skill does not contain any scripts, binaries, or dependencies. It relies entirely on natural language instructions to guide the agent's behavior, which significantly limits the attack surface.
[INDIRECT_PROMPT_INJECTION]: The skill is designed to process user-provided decisions. While there are no explicit delimiters to separate the 'decision' from the instructions, the skill possesses no capabilities (such as file-writing or network access) that could be exploited even if a user were to embed instructions within their decision. The impact of such an injection would be limited to the content of the generated report.

dissent