context-degradation
Pass
Audited by Gen Agent Trust Hub on Jun 30, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill's primary purpose is educational and diagnostic. The provided Python script (
scripts/degradation_detector.py) uses standard libraries (random, re, typing) to perform statistical analysis and pattern matching on provided text strings to simulate attention patterns and detect potential errors or contradictions. - [COMMAND_EXECUTION]: The skill does not contain any instructions that attempt to execute shell commands or interact with the underlying operating system in a privileged way. The code snippets in the documentation and the scripts are restricted to text processing and analysis.
- [DATA_EXFILTRATION]: No network operations (curl, wget, fetch) or remote data transmissions were detected. The skill operates entirely on data provided to its functions locally.
- [PROMPT_INJECTION]: The instructions in SKILL.md and technical references focus on assisting the agent with context management and do not contain any instructions to bypass safety filters or override system constraints.
- [REMOTE_CODE_EXECUTION]: There are no external package requirements (pip/npm) or remote script downloads. All logic is implemented within the provided local files.
Audit Metadata