create-git-issue

Pass

Audited by Gen Agent Trust Hub on Jul 18, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTIONEXTERNAL_DOWNLOADS
Full Analysis
  • [PROMPT_INJECTION]: The "Skill Isolation" section includes instructions for the agent to suppress other skills and disregard external interruptions, which attempts to override standard orchestration and behavioral controls.
  • Evidence: "No other skill may activate, interrupt, or modify this skill's behavior... suppress it and continue without interruption."
  • [COMMAND_EXECUTION]: The skill uses shell commands via the GitHub CLI (gh) to create repository issues, using content generated from ingested external data.
  • Evidence: gh issue create --title "<PRD title>" --body-file <prd-body-file>.md.
  • Ingestion points: The skill reads from technical_requirements.md and content from external GitHub issue bodies and comments.
  • Boundary markers: No explicit delimiters or boundary markers are used for external data interpolation.
  • Capability inventory: Shell command execution via gh CLI and file system write operations (prd.md, issues.md).
  • Sanitization: Employs mandatory user approval steps for both the synthesized PRD and the implementation slice breakdown before execution.
  • [EXTERNAL_DOWNLOADS]: The skill interacts with GitHub's public services to fetch issue context and publish new documentation.
  • Evidence: Fetches information and creates issues via the official GitHub CLI tool.
Audit Metadata
Risk Level
SAFE
Analyzed
Jul 18, 2026, 10:56 AM
Security Audit — agent-trust-hub — create-git-issue