rust-gpu-discipline

Pass

Audited by Gen Agent Trust Hub on Jul 9, 2026

Risk Level: SAFECOMMAND_EXECUTION
Full Analysis
  • [SAFE]: The skill provides instructions for the agent to avoid common coding anti-patterns when writing GPU-accelerated Rust code. These directives improve implementation honesty and do not attempt to bypass security guardrails or safety filters.\n- [COMMAND_EXECUTION]: The skill instructs the agent to run local system probes such as nvidia-smi, ls /usr/local/cuda*, and grep on workspace configuration files to detect hardware and toolkit availability. It also specifies the use of cargo build and cargo test for local verification. These commands are standard for development environments and do not involve network-based exfiltration or privilege escalation.\n- [EXTERNAL_DOWNLOADS]: URL references to GitHub (forecast-bio/ferray) and TomeVault (tomevault.io) are used for attribution and documentation of the skill's source and distribution. No executable scripts or third-party binaries are downloaded for execution.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 9, 2026, 01:22 AM
Security Audit — agent-trust-hub — rust-gpu-discipline