git-onboarding-auto
Pass
Audited by Gen Agent Trust Hub on Mar 3, 2026
Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSREMOTE_CODE_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes extensive bash commands to collect system state and perform Git operations, including modifying global configurations and managing repository remotes.
- [EXTERNAL_DOWNLOADS]: Fetches and installs the Homebrew package manager and the GitHub CLI (gh) from their official repositories to fulfill environment prerequisites.
- [REMOTE_CODE_EXECUTION]: Downloads and executes the official Homebrew installation script using a shell pipe (curl | bash). While this originates from a well-known service, it constitutes remote code execution at runtime.
- [PROMPT_INJECTION]: The skill has a vulnerability surface for indirect prompt injection (Category 8).
- Ingestion points: The agent ingests and summarizes output from local repository commands such as
git status,git remote get-url, andgit log @{u}..in Phase 1. - Boundary markers: There are no delimiters or instructions provided to the agent to ignore potentially malicious content embedded in the repository's metadata or commit history.
- Capability inventory: The agent possesses the ability to execute arbitrary bash commands, write files, and interact with network APIs via the GitHub CLI.
- Sanitization: No sanitization or filtering is performed on the data retrieved from the local repository environment before it is passed to the agent for processing.
Audit Metadata