github-create-issues

Pass

Audited by Gen Agent Trust Hub on Aug 30, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTIONPRIVILEGE_ESCALATION
Full Analysis
  • [COMMAND_EXECUTION]: The create_github_issues.py script uses subprocess.run to call the GitHub CLI (gh auth token) to retrieve authentication credentials. This is a functional requirement for interacting with GitHub and does not involve arbitrary shell execution.
  • [EXTERNAL_DOWNLOADS]: The skill documentation instructs the installation of requests and openpyxl via pip. These are widely-used, reputable packages required for network operations and data processing.
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests untrusted data from github_backlog_input.json (ingestion point) and uses it to create GitHub milestones, labels, and issues using the requests library (capability inventory). The skill implements a safety gate requiring a visual dry-run table for user review and explicit approval (boundary markers) before any API calls are executed. No specific sanitization is performed on the input data before processing.
  • [PRIVILEGE_ESCALATION]: The SKILL.md file suggests executing a setup verification script with the PowerShell -ExecutionPolicy Bypass flag. While this bypasses default execution restrictions, the referenced script is a diagnostic tool that checks prerequisites without modifying system state.
Audit Metadata
Risk Level
SAFE
Analyzed
Aug 30, 2026, 02:27 AM
Security Audit — agent-trust-hub — github-create-issues