prompt-enhance
Warn
Audited by Gen Agent Trust Hub on Oct 1, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTIONDYNAMIC_EXECUTION
Full Analysis
- [COMMAND_EXECUTION]: The skill uses a shell command to interact with a prompt database: `python3 {PROMPT_ENGINE_DIR}/scripts/search_prompts.py "KEY_TERMS" --limit 5".\n- [INDIRECT_PROMPT_INJECTION]: The skill has a surface for indirect prompt injection as it processes untrusted user-supplied prompts.\n
- Ingestion points: User prompts provided for enhancement.\n
- Boundary markers: The instructions do not define boundary markers (e.g., delimiters) to separate user content from agent instructions.\n
- Capability inventory: The skill has the capability to execute shell commands (
python3).\n - Sanitization: There is no mention of sanitizing or escaping the "KEY_TERMS" extracted from the user prompt before passing it to the shell command.\n- [DYNAMIC_EXECUTION]: The skill performs dynamic execution by running a Python script with arguments generated at runtime from the conversation flow.
Audit Metadata