The Agent Skills Directory

[SAFE]: The skill is entirely documentation-based and does not include any scripts, binary files, or executable instructions. It functions as a logic framework for the AI to follow when analyzing experiment data provided by the user.
[DATA_EXFILTRATION]: No network-capable tools (such as curl, wget, or fetch) or sensitive file access patterns were detected. The skill does not request access to credentials, environment variables, or local configuration files.
[REMOTE_CODE_EXECUTION]: There are no patterns involving the download or execution of remote code. All referenced URLs are static links to reputable documentation and research papers from well-known organizations (Microsoft, Statsig, Nielsen Norman Group).
[PROMPT_INJECTION]: The instructions do not contain attempts to override safety filters, bypass system prompts, or adopt unrestricted personas. The language is professional and focused on the stated domain of statistical analysis.
[OBFUSCATION]: No hidden content, Base64 encoding, homoglyphs, or zero-width characters were found in any of the skill files.
[INDIRECT_PROMPT_INJECTION]: While the skill is designed to process external data (experiment results), it does not automate actions or use tools based on that data. It serves as an interpretive layer, which minimizes the risk of indirect injection becoming an attack vector.

experiment-results-interpreter