The Agent Skills Directory

[SAFE]: The skill performs an audit of context window consumption by reading local files such as agents, skills, rules, and configuration files like .mcp.json and GEMINI.md. This behavior is consistent with its stated purpose.
[SAFE]: No network activity, data exfiltration, or unauthorized command execution was identified. The skill operates entirely within the local environment for analysis.
[SAFE]: The logic focuses on token estimation and reporting, providing actionable recommendations to reclaim context space without bypassing safety filters or security protocols.

context-budget