The Agent Skills Directory

[SAFE]: The skill is a purely instructional template for logical reasoning and documentation. It defines a structured way to handle claims and counterclaims without involving any technical operations.
[NO_CODE]: No scripts, binaries, or executable commands are included in the skill definition. The provided procedure consists of logical steps for the AI to follow in natural language.
[INDIRECT_PROMPT_INJECTION]: The skill processes external input (focal_claim, counterclaim, evidence_records). While these are untrusted data sources, the skill lacks any 'allowed-tools' or capabilities to execute shell commands, perform network requests, or modify the file system. Consequently, even if a prompt injection was present in the data, it would have no mechanism to cause harm beyond influencing the text output of this specific reasoning task.

document-counterclaim