impeccable
Pass
Audited by Gen Agent Trust Hub on Oct 30, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes local Python scripts for state inspection and findings triage. These scripts (doctor.py, triage_findings.py) are implemented using only the Python standard library and operate in a read-only manner without network access or external command execution.
- [EXTERNAL_DOWNLOADS]: The skill documents that the upstream engine may download its binary components from its official GitHub repository (pbakaus/impeccable). These operations are described as gated and require explicit user approval, with the skill providing integrity verification via digest checks.
- [INDIRECT_PROMPT_INJECTION]: The skill processes data from a design detector that ingests external UI code. It contains explicit instructions for the agent to treat snippets from these files as untrusted data and includes a triage script designed to exclude these snippets from its output to mitigate potential injection risks.
Audit Metadata