ljg-word-flow

Pass

Audited by Gen Agent Trust Hub on Mar 27, 2026

Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
  • [PROMPT_INJECTION]: The skill includes instructions to 'Force NATIVE mode' and explicitly bypass the agent's standard reasoning process (OBSERVE/THINK/PLAN/BUILD/EXECUTE/VERIFY/LEARN). This attempt to override the agent's default operational loop and safety/planning protocols is a behavioral injection pattern.\n- [COMMAND_EXECUTION]: The workflow involves executing local command-line tools ljg-word and ljg-card -i. It further interacts with the file system by generating and saving PNG files into the ~/Downloads/ directory.
Audit Metadata
Risk Level
SAFE
Analyzed
Mar 27, 2026, 04:21 PM