evaluate-waza-copilot-customizations

Pass

Audited by Gen Agent Trust Hub on Nov 11, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [COMMAND_EXECUTION]: The skill frequently invokes the waza CLI and executes local PowerShell scripts (e.g., grade-copilot-customizations.ps1, write-runtime-task.ps1) to perform its evaluation and reporting tasks.
  • [DATA_EXPOSURE]: The skill is designed to read from sensitive local directories, specifically $HOME/.copilot, to evaluate personal Copilot instructions and customizations. While this is the intended purpose, users should be aware that the agent accesses these personalization files.
  • [INDIRECT_PROMPT_INJECTION]: The skill processes untrusted data in the form of Copilot customization files (instructions, skills, agents) which could contain embedded malicious instructions intended to influence the evaluation agent.
  • Ingestion points: Customization files located in ~/.copilot, <workspace>/.github, or user-specified paths.
  • Boundary markers: No explicit delimiters or boundary markers are defined to isolate the evaluated content from the skill's operational instructions.
  • Capability inventory: PowerShell subprocess execution, file system read/write operations, and interaction with the waza CLI tool.
  • Sanitization: The skill instructions recommend manually excluding secrets from final reports, but there is no automated sanitization of the input data processed during evaluation.
  • [EXTERNAL_DOWNLOADS]: The skill references external setup guides and templates hosted on the author's GitHub repository (github.com/aktsmm/Agent-Customization) for initializing the evaluation environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Nov 11, 2026, 04:17 AM