vet-flow-guide
Pass
Audited by Gen Agent Trust Hub on Sep 8, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill defines a read-only process for comparing implementation against specs. It explicitly forbids mutating product code or the run file during the judgment phase.
- [REMOTE_CODE_EXECUTION]: No remote script downloads, untrusted package installations, or command execution from external sources were identified. The skill uses platform-native subagents for task isolation.
- [DATA_EXFILTRATION]: The skill does not perform network operations or access sensitive system directories such as .ssh or .aws. File operations are limited to project-specific paths.
- [COMMAND_EXECUTION]: No shell command execution or subprocess spawning instructions are present in the skill body.
- [OBFUSCATION]: No encoded strings, hidden characters, or homoglyph-based URL variations were detected in the skill metadata or instructions.
Audit Metadata