prompt-build

Pass

Audited by Gen Agent Trust Hub on Oct 31, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill instructions direct the agent to execute a shell command that incorporates user-supplied text directly as a command-line argument.
  • Evidence: python3 {PROMPT_ENGINE_DIR}/scripts/search_prompts.py "SUBJECT" --model MODEL --limit 5 in SKILL.md.
  • [INDIRECT_PROMPT_INJECTION]: The skill possesses a surface for indirect injection by processing untrusted user input and using it to drive script execution and prompt generation without explicit sanitization.
  • Ingestion points: User-provided values for 'Subject', 'Target model', and 'Style/mood' defined in Step 1 of SKILL.md.
  • Boundary markers: None are specified to delimit user input when it is interpolated into the shell command or the final prompt structure.
  • Capability inventory: File reading (model-guide.md) and local script execution (search_prompts.py).
  • Sanitization: No instructions are provided to sanitize or validate the 'SUBJECT' or 'MODEL' variables before they are passed to the shell environment.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 31, 2026, 05:14 PM