skills/firzus/agent-skills/imagegen/Gen Agent Trust Hub

imagegen

Pass

Audited by Gen Agent Trust Hub on May 12, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSDATA_EXFILTRATION
Full Analysis
  • [COMMAND_EXECUTION]: The skill utilizes a bash wrapper script to invoke the local codex CLI and auxiliary Python scripts. It uses bash arrays and heredocs to safely handle user-provided prompts, preventing shell injection vulnerabilities.
  • [EXTERNAL_DOWNLOADS]: The skill depends on the codex CLI (OpenAI) and the Pillow Python library. These are well-known, trusted tools from established providers.
  • [DATA_EXFILTRATION]: The skill reads image data from the user's local ~/.codex/sessions/ directory. It implements a snapshotting and diffing mechanism to ensure it only processes session files generated during its current execution, protecting the privacy of unrelated user conversations.
Audit Metadata
Risk Level
SAFE
Analyzed
May 12, 2026, 09:39 PM
Security Audit — agent-trust-hub — imagegen