sre-engineer

Fail

Audited by Gen Agent Trust Hub on Jul 15, 2026

Risk Level: HIGHCOMMAND_EXECUTIONREMOTE_CODE_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: Multiple scripts execute system-level commands using subprocess.run. Examples include using kubectl to restart deployments or delete pods, systemctl to restart services, and tc or iptables for network manipulation in chaos experiments.
  • [REMOTE_CODE_EXECUTION]: The toil automation script in SKILL.md implements a pattern where data fetched via urllib.request from a Prometheus API (http://prometheus:9090) directly determines whether a kubectl rollout restart command is executed. This establishes a remote data-to-command execution flow.
  • [COMMAND_EXECUTION]: The AutomatedRunbook class in references/automation-toil.md utilizes subprocess.run with shell=True. Executing arbitrary strings as shell commands increases the risk of command injection if the input commands are not strictly validated.
  • [EXTERNAL_DOWNLOADS]: The skill initiates network requests using urllib.request.urlopen (in SKILL.md) and curl (in references/automation-toil.md) to fetch metrics and perform health checks from external or local endpoints.
Recommendations
  • HIGH: Downloads and executes remote code from: http://prometheus:9090/api/v1/query?query={urllib.request.quote(query)} - DO NOT USE without thorough review
Audit Metadata
Risk Level
HIGH
Analyzed
Jul 15, 2026, 04:40 PM
Security Audit — agent-trust-hub — sre-engineer