Context Manager
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: SAFENO_CODE
Full Analysis
- [NO_CODE] (SAFE): The skill consists entirely of Markdown documentation (SKILL.md) providing strategies and templates for context management. It does not include any Python, JavaScript, or shell scripts.
- [EXTERNAL_DOWNLOADS] (SAFE): No external package manifests (package.json, requirements.txt) or remote script download commands (curl, wget) were detected.
- [DATA_EXFILTRATION] (SAFE): The skill does not contain any network operations or logic to access sensitive local files. It focuses on managing the LLM's internal context window.
- [PROMPT_INJECTION] (SAFE): The instructions are benign and focused on conversation optimization. No attempts to bypass safety filters or override system instructions were found.
- [COMMAND_EXECUTION] (SAFE): No subprocess spawning or shell command execution logic is present.
Audit Metadata