The Agent Skills Directory

[SAFE]: The skill implements context management logic using standard Python libraries. No dangerous functions such as eval, exec, or subprocess are present.
[SAFE]: No network operations, sensitive file access, or hardcoded credentials were detected. Data processing is restricted to local string manipulation and regex-based extraction.
[SAFE]: The instructions and scripts focus on legitimate context optimization techniques like KV-cache stabilization and observation masking. No patterns of prompt injection or persistence mechanisms were found.
[SAFE]: Metadata and documentation accurately reflect the skill's purpose and functionality.

context-optimization