The Agent Skills Directory

[SAFE]: The skill consists of natural language instructions intended to guide the agent's internal reasoning process. It does not include any executable code, shell scripts, or tool invocations.
[INDIRECT_PROMPT_INJECTION]: The skill ingests user-supplied content via the $ARGUMENTS variable in SKILL.md. While it does not implement specific delimiters or sanitization for this input, the skill lacks any capabilities (such as file system access, network requests, or command execution) that could be leveraged by a malicious payload.

calibrate-confidence