The Agent Skills Directory

[SAFE]: The skill's workflow and instructions are focused on legitimate development tasks such as code implementation, review, and documentation. It does not contain any suspicious commands, obfuscation, or data exfiltration patterns.
[DATA_EXFILTRATION]: The skill references internal documentation files using relative directory traversal (e.g., ../../../70-knowledge/). This access is limited to internal documentation within the project workspace and does not target sensitive system configuration files, environment variables, or user credentials. No network operations are present to exfiltrate data.

on-spring