sn-infographic

Pass

Audited by Gen Agent Trust Hub on May 27, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface by ingesting untrusted data through the user_prompt parameter. This input is interpolated into system prompts during the content analysis (Step 2.0) and prompt expansion (Step 2.3) phases without boundary markers or explicit instructions to disregard embedded commands. This could allow an attacker to influence the downstream generation logic or exfiltrate processed content.
  • [COMMAND_EXECUTION]: The workflow defined in SKILL.md (specifically Steps 1, 2.0, 2.3, and 3) instructs the agent to construct and execute shell commands using variables derived directly from user input, such as $USER_PROMPT and $STRUCTURED_CONTENT. If the execution environment does not properly escape these shell variables, it creates a significant surface for command injection attacks.
  • [DATA_EXPOSURE]: The skill uses a hardcoded temporary directory path /tmp/openclaw/sn-infographic/ for storing task-related artifacts, including analysis results and generated images. This could lead to data exposure if the directory permissions are not strictly managed or if task IDs are predictable.
Audit Metadata
Risk Level
SAFE
Analyzed
May 27, 2026, 06:44 AM
Security Audit — agent-trust-hub — sn-infographic