modeless-ui-design
Pass
Audited by Gen Agent Trust Hub on Sep 23, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill instructs the agent to execute a bundled Python script,
scripts/validate-review.py, to verify that the generated UI review document adheres to the required structure. This is a local execution of a provided utility script that uses standard libraries for pattern matching and does not perform network operations. - [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted user input (UI requirements and user flows) and subsequently executes a shell command involving the generated output path. This creates a surface where malicious input could attempt to influence the execution environment.
- Ingestion points: User-provided descriptions of UI screens, objects, and user flows in SKILL.md Step 1.
- Boundary markers: Absent.
- Capability inventory: Local execution of
python3 scripts/validate-review.py <成果物パス>in SKILL.md Step 8. - Sanitization: Absent; the validation script relies on strict heading matches but does not sanitize the input content itself.
Audit Metadata