architecture
Pass
Audited by Gen Agent Trust Hub on Feb 17, 2026
Risk Level: LOWNO_CODE
Full Analysis
- [NO_CODE] (SAFE): The skill consists exclusively of Markdown documentation and logic templates for architectural design. No scripts, binaries, or configuration files are present.
- [EXTERNAL_DOWNLOADS] (SAFE): No external resources, libraries, or scripts are downloaded or referenced for runtime use.
- [COMMAND_EXECUTION] (SAFE): There are no shell commands, system calls, or subprocess executions defined within the skill files.
- [DATA_EXFILTRATION] (SAFE): The skill does not access local sensitive files or perform any network communication.
- [PROMPT_INJECTION] (SAFE): Instructions are domain-specific for architecture and do not attempt to subvert agent behavior or extract system prompts.
- [INDIRECT_PROMPT_INJECTION] (SAFE): While the skill gathers project context from the user, it lacks any side-effect capabilities (write, execute, network) to be exploited via injected content.
Audit Metadata