skills/cursor/plugins/arena/Gen Agent Trust Hub

arena

Pass

Audited by Gen Agent Trust Hub on Jun 13, 2026

Risk Level: SAFEPROMPT_INJECTION
Full Analysis
  • [SAFE]: The skill provides a structured framework for managing parallel AI agent tasks. It utilizes isolated temporary directories for file storage and standard platform-native subagent management features for task execution.
  • [PROMPT_INJECTION]: The orchestration workflow involves reading and merging content from multiple subagents, which represents a surface for indirect prompt injection.
  • Ingestion points: Phase D (reading candidates end-to-end) and Phase E (identifying grafts from losing candidates) in SKILL.md.
  • Boundary markers: Absent; the skill instructions do not specify the use of delimiters or instructions to ignore embedded prompts when processing subagent outputs.
  • Capability inventory: The skill enables high-level synthesis and writing to designated file paths but does not explicitly invoke dangerous shell commands or network operations.
  • Sanitization: Absent; candidate outputs are merged into the final artifact without explicit validation or filtering of their content.
Audit Metadata
Risk Level
SAFE
Analyzed
Jun 13, 2026, 12:38 AM
Security Audit — agent-trust-hub — arena