glab-issue

Pass

Audited by Gen Agent Trust Hub on Jun 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill performs shell command execution using 'glab' and 'git' to interact with GitLab services and local repositories.
  • The script scripts/batch-label-issues.sh executes glab issue update in a loop over provided issue IDs.
  • The script scripts/create-mr-from-issue.sh executes glab, git checkout, git commit, and git push to automate the transition from an issue to a merge request.
  • [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection through issue content fetched from GitLab.
  • Ingestion points: In scripts/create-mr-from-issue.sh, the variable ISSUE_TITLE is populated with content from an external GitLab issue title via glab issue view.
  • Capability inventory: The skill has the capability to write to the local filesystem (git branch creation), create commits (git commit), and perform network operations (git push, glab mr create).
  • Sanitization: The script performs sanitization on the issue title before using it in a branch name (converting to lowercase and replacing non-alphanumeric characters with hyphens). However, the raw title is still used in the commit message and terminal output.
  • Boundary markers: The author has included an explicit warning comment in scripts/create-mr-from-issue.sh stating that the issue title is untrusted and should not be executed or evaluated.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 22, 2026, 05:22 AM
Security Audit — agent-trust-hub — glab-issue