parallel-agents
Pass
Audited by Gen Agent Trust Hub on Feb 16, 2026
Risk Level: LOW
Full Analysis
- [SAFE] (SAFE): The skill provides purely instructional content for agent orchestration. It does not contain any executable scripts, obfuscated payloads, or prompt injection attempts.
- [COMMAND_EXECUTION] (SAFE): The 'allowed-tools' section is strictly limited to 'Read', 'Glob', and 'Grep', which are non-destructive, read-only operations that do not permit arbitrary shell access or file modification.
- [DATA_EXFILTRATION] (SAFE): No network-enabled tools (e.g., curl, wget) are requested, and no external domains are specified for data transmission, ensuring all analysis remains local.
- [REMOTE_CODE_EXECUTION] (SAFE): There are no patterns involving remote downloads or dynamic execution of untrusted external content.
Audit Metadata