skills/tangc/tangzhan-skills/jimeng/Gen Agent Trust Hub

jimeng

Pass

Audited by Gen Agent Trust Hub on Oct 8, 2026

Risk Level: SAFECOMMAND_EXECUTIONEXTERNAL_DOWNLOADSINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The shell script jimeng.sh invokes a Python helper sign.py to perform HMAC-SHA256 signing for Volcengine API requests. This is a functional requirement and uses local files included in the skill.
  • [EXTERNAL_DOWNLOADS]: The skill performs network operations via curl to the Volcengine API (visual.volcengineapi.com) and a local bridge for Feishu (localhost:18789). These are documented behaviors for image generation and notification.
  • [INDIRECT_PROMPT_INJECTION]: The skill uses shell variable interpolation to construct JSON payloads for the Volcengine API. It lacks sanitization for characters like double quotes in the user-supplied prompt, which could allow for JSON structure manipulation if the agent processes untrusted data.
  • Ingestion points: The prompt and reference_url parameters in jimeng.sh.
  • Boundary markers: None.
  • Capability inventory: HTTP POST requests via curl to the Volcengine API and a local message endpoint.
  • Sanitization: No escaping or validation is performed on the prompt string before JSON construction.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 8, 2026, 10:34 PM