codebase-onboarding

Pass

Audited by Gen Agent Trust Hub on Jul 20, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes common development and system tools such as git, gh (GitHub CLI), docker, and grep to map codebase architecture and state. These are used for reconnaissance as part of the intended functionality.
  • [COMMAND_EXECUTION]: It encourages running introspection commands against locally installed libraries to verify versions and structures, which involves executing code in the local environment for auditing purposes.
  • [PROMPT_INJECTION]: The skill processes external codebase content such as READMEs and code files. While this presents a surface for indirect prompt injection, the skill includes explicit instructions for the agent to remain in a 'READ-ONLY' state and to verify facts independently, mitigating risk.
  • [SAFE]: Explicit rules are provided to prevent modification of the system or codebase during the onboarding phase, minimizing the risk of accidental damage or unauthorized persistence.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 20, 2026, 07:18 AM
Security Audit — agent-trust-hub — codebase-onboarding