The Agent Skills Directory

[SAFE]: The skill consists entirely of natural language instructions for reasoning and text analysis. No scripts, binaries, or executable code are included.
[SAFE]: The skill ingests user input through the $ARGUMENTS variable in SKILL.md to identify claims for calibration. While there are no explicit delimiters or sanitization routines, the skill has no access to dangerous tools (file system, network, or shell) that could be exploited via indirect prompt injection.
[SAFE]: External references in README.md point to the author's own GitHub repository for installation and licensing, which is standard practice for skill distribution.

calibrate-confidence