install-cognitive-os
Pass
Audited by Gen Agent Trust Hub on Jun 17, 2026
Risk Level: SAFE
Full Analysis
- [COMMAND_EXECUTION]: The skill performs file system operations to write configuration files for various AI-integrated tools, including VS Code, Cursor, and Claude Code.- [PROMPT_INJECTION]: The skill serves as a mechanism for propagating operational instructions across different agent platforms. While this is the intended behavior of the installer, it establishes an indirect prompt injection surface by bridging rules between agent contexts.
- Ingestion points: Retrieves source cognitive rules from
STYLE.mdor the agent's memory system. - Boundary markers: The skill does not explicitly define or use isolation delimiters for the instructions it installs into target files.
- Capability inventory: Broad file-write access is used to update environment-specific configuration paths (SKILL.md).
- Sanitization: There is no specific validation or sanitization step for the 'Cognitive Loop' content before it is extracted and installed.
Audit Metadata