context-pruning-directive
Pass
Audited by Gen Agent Trust Hub on Oct 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill implements a performance-optimization utility for context management. The provided Python middleware is a standard data transformation function that truncates long strings in a message history array without using unsafe built-in functions or system calls.
- [SAFE]: No network operations, credential exposures, or sensitive file access patterns were detected. The skill operates entirely on local data structures provided to its functions.
- [SAFE]: The 'Agent Operational Directive' is a technical instruction aimed at improving token efficiency and agent attention; it does not contain attempts to bypass safety filters, extract system prompts, or override core behavioral guidelines.
- [SAFE]: No obfuscation, remote code execution, or persistence mechanisms were found in the instructions or the code snippet.
Audit Metadata