nemo-relay-plugin-build

Pass

Audited by Gen Agent Trust Hub on Aug 13, 2026

Risk Level: SAFE
Full Analysis
  • [SAFE]: The skill focuses on instructing agents how to package reusable runtime behavior for NeMo Relay. It explicitly mandates that configuration validation must be deterministic and must not perform side-effect-heavy operations like opening network connections or mutating process state.
  • [INDIRECT_PROMPT_INJECTION]: The skill describes a surface for processing external configuration data.
  • Ingestion points: Plugin configuration JSON documents (SKILL.md).
  • Boundary markers: Not explicitly defined for variable interpolation.
  • Capability inventory: The skill allows registering runtime behavior including subscribers, guardrails, and intercepts (SKILL.md).
  • Sanitization: The skill includes mandatory instructions to define diagnostics for unsafe configuration and stop execution before registration if validation fails (SKILL.md).
  • [SAFE]: Dependencies such as 'nemo_relay' and 'nemo-relay-node' are official vendor components consistent with the NVIDIA NeMo framework.
  • [SAFE]: The presence of comprehensive evaluation benchmarks (BENCHMARK.md) and test cases (evals/evals.json) indicates that security and behavior checks were performed during the skill's development.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 13, 2026, 07:21 PM
Security Audit — agent-trust-hub — nemo-relay-plugin-build