context-degradation
Pass
Audited by Gen Agent Trust Hub on Oct 29, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [SAFE]: The skill is a documentation-only resource focused on educational content regarding LLM context degradation patterns.
- [NO_CODE]: There are no executable scripts, shell commands, or dynamic execution patterns present in the file.
- [EXTERNAL_DOWNLOADS]: The skill includes links to academic research papers on arxiv.org (Lost in the Middle, RULER). These are standard references for the subject matter and do not represent a security risk.
- [PROMPT_INJECTION]: No patterns of instruction overriding, safety bypasses, or adversarial roleplay instructions were detected.
Audit Metadata