nvflare-diagnose-job

Pass

Audited by Gen Agent Trust Hub on Jul 6, 2026

Risk Level: SAFE
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes the official nvflare command-line interface to perform diagnostic tasks. These operations are strictly limited to read-only functions such as retrieving job metadata, logs, statistics, and system status.
  • [DATA_EXFILTRATION]: The instructions explicitly forbid reading private key contents and emphasize collecting bounded logs and job artifacts to prevent excessive data exposure. No unauthorized network transmission or exfiltration patterns were identified.
  • [PROMPT_INJECTION]: The instructions are technical and focused on interpreting NVFLARE failure patterns. They do not contain any attempts to bypass safety filters, extract system prompts, or override agent constraints.
  • [SAFE]: The skill follows security best practices by specifying a read_only blast radius and providing a predefined catalog of failure patterns for systematic diagnosis.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 6, 2026, 05:37 AM
Security Audit — agent-trust-hub — nvflare-diagnose-job