audit-architecture
Pass
Audited by Gen Agent Trust Hub on Aug 14, 2026
Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes shell commands to analyze the repository structure and content.
- Subagent 8 uses a subshell pattern:
wc -l $(rg -l '...' packages/*/src/ 2>/dev/null). This construction is vulnerable to command injection if a file in the repository has a name containing shell metacharacters (e.g.,;,&, or backticks), as the shell may execute the filename as a command. - [SAFE]: The skill's primary purpose is a read-only audit of architectural drift, which is a legitimate development workflow tool.
- The subagents are explicitly defined as read-only and focused on exploration (
subagent_type=Explore). - There are no detected hardcoded credentials, obfuscated payloads, or attempts at data exfiltration to external domains.
- All referenced resources and naming conventions align with the stated vendor, Metabase.
Audit Metadata