skills/luongnv89/skills/system-design/Gen Agent Trust Hub

system-design

Pass

Audited by Gen Agent Trust Hub on Apr 22, 2026

Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill performs automated Git repository synchronization tasks to ensure a clean working tree and remote alignment. Evidence: uses git fetch, git pull --rebase, and git push in SKILL.md to manage branch states during the workflow.
  • [COMMAND_EXECUTION]: The skill optionally executes a localized Python script for project maintenance. Evidence: Phase 6 in SKILL.md attempts to run python3 scripts/update_readme_ideas_index.py if the script exists in the local project directory.
  • [PROMPT_INJECTION]: The skill possesses an indirect prompt injection surface as it processes untrusted user-provided Markdown content to drive core logic. Evidence:
  • Ingestion points: Reads prd.md, idea.md, and validate.md (via prd-reader.md).
  • Boundary markers: Absent; extracted requirements are interpolated directly into research prompts.
  • Capability inventory: Includes repository write access, Git push capability, and local script execution.
  • Sanitization: None; input from documents is used to define architecture decisions and technical choices.
Audit Metadata
Risk Level
SAFE
Analyzed
Apr 22, 2026, 04:13 AM
Security Audit — agent-trust-hub — system-design