The Agent Skills Directory

[SAFE]: No malicious patterns, exfiltration attempts, or unsafe commands were found. The skill follows best practices for OKR evaluation and maintains clear boundaries for data interpretation.
[INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user data (OKR values, targets, evidence). However, it lacks dangerous capabilities like subprocess execution, file system writes, or network operations that could be exploited via injection. It also mandates validation of inputs before grading.

measure-okr-grader