performance-engineer

Pass

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill guides the agent to use high-privilege system diagnostic and configuration utilities such as perf, tcpdump, and sysctl. These tools are required for kernel-level tuning and system-wide profiling but provide extensive access to system state and the ability to modify kernel parameters.
  • [INDIRECT_PROMPT_INJECTION]: The skill involves processing and analyzing external data sources, creating a potential surface for indirect prompt injection.
  • Ingestion points: The agent is instructed to read and analyze performance stack traces, profile files (e.g., out.perf), and application latency metrics (SKILL.md).
  • Boundary markers: The workflows do not define specific delimiters or instructions to ignore potential commands embedded within analyzed log or profile data.
  • Capability inventory: The agent utilizes powerful capabilities including system-wide performance recording (perf record), network traffic capture (tcpdump), and kernel parameter modification (sysctl).
  • Sanitization: There are no explicit validation or sanitization steps mentioned for the external data ingested during profiling and analysis workflows.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 23, 2026, 08:05 AM
Security Audit — agent-trust-hub — performance-engineer