design-invariants
Pass
Audited by Gen Agent Trust Hub on May 8, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill provides numerous shell-based diagnostics using tools like
rg(ripgrep),find,grep, andbash. These are used to identify architectural violations within the repository. Notable commands include executing a local audit script and runningnpmcommands for validation. - [PROMPT_INJECTION]: The skill exhibits a surface for indirect prompt injection by processing untrusted data from design proposals, pull request diffs, and URLs. Malicious instructions could be embedded in these artifacts to influence the agent's audit results.
- Ingestion points: User-provided design paths, diff ranges, and PR URLs.
- Boundary markers: None identified in the prompt templates or instructions.
- Capability inventory: Access to local repository files and shell command execution (
rg,find,bash,npm). - Sanitization: No input sanitization or validation logic is defined for the audited content.
Audit Metadata