k8s-maestro-integrator

Pass

Audited by Gen Agent Trust Hub on Apr 1, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The skill contains logic for generating code from user descriptions, which constitutes a surface for indirect prompt injection.\n
  • Ingestion points: User-provided strings are passed as command line arguments to scripts/generate_workflow_code.py and scripts/generate_test_code.py.\n
  • Boundary markers: None. Input is processed directly from CLI arguments without delimiters or safety instructions.\n
  • Capability inventory: The skill possesses capabilities for filesystem writes and shell command execution (via kubectl in related scripts).\n
  • Sanitization: The scripts employ keyword matching to select hardcoded templates, which limits the influence of user input on the final output to template selection.\n- [COMMAND_EXECUTION]: Multiple scripts, including analyze_cluster.sh and detect_crate_features.py, execute kubectl commands to gather cluster metadata and recommend crate features. These operations use hardcoded arguments and do not present a command injection risk.\n- [EXTERNAL_DOWNLOADS]: Installation instructions reference the author's official GitHub repository. No external code is dynamically downloaded or executed at runtime beyond these manual installation steps.\n- [SAFE]: The skill demonstrates security best practices by providing tutorials on managing Kubernetes Secrets and avoiding hardcoded credentials.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 1, 2026, 05:07 PM
Security Audit — agent-trust-hub — k8s-maestro-integrator