skills/raine/workmux/rebase/Gen Agent Trust Hub

rebase

Pass

Audited by Gen Agent Trust Hub on Sep 14, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill makes extensive use of the Bash tool to execute shell commands such as git status, git stash, git fetch, git rebase, and git log. These commands are used to automate the git workflow.
  • [INDIRECT_PROMPT_INJECTION]: The skill is vulnerable to command injection through the processing of untrusted user data.
  • Ingestion points: The skill accepts user-provided input via the $ARGUMENTS variable in SKILL.md.
  • Boundary markers: There are no delimiters or boundary markers used to separate user input from the shell command structure.
  • Capability inventory: The skill uses the Bash tool to execute commands based on the parsed $ARGUMENTS.
  • Sanitization: No sanitization or validation logic is defined to check the $ARGUMENTS for shell metacharacters (e.g., ;, &&, |, or backticks), creating a surface for command injection during the git rebase <target> or git fetch <remote> steps.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 14, 2026, 09:14 PM
Security Audit — agent-trust-hub — rebase