The Agent Skills Directory

[SAFE]: The skill serves as a legitimate technical resource for context management and engineering. A thorough manual review of the provided instructions, Python scripts, and reference documents found no evidence of malicious intent, credential exposure, or unauthorized system access.
[PROMPT_INJECTION]: The skill is designed to ingest and analyze external conversation history, which represents an indirect prompt injection surface. However, the implementation is restricted to string manipulation for summarization and quality evaluation, and the skill does not utilize high-privilege tools or system calls (such as file writes or network exfiltration) that would allow an attacker to exploit the processed data.

context-compression