prompt-adapt

Pass

Audited by Gen Agent Trust Hub on Oct 31, 2026

Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
  • [COMMAND_EXECUTION]: The skill uses a shell command to execute a local Python script "search_prompts.py" to find reference prompts based on user input for a target model.\n- [INDIRECT_PROMPT_INJECTION]: The skill is susceptible to command injection because it interpolates user-controlled data ("SUBJECT" and "TARGET_MODEL") directly into a shell command line without specifying sanitization or validation protocols.\n
  • Ingestion points: User-supplied prompt content and model names provided during the adaptation workflow in SKILL.md.\n
  • Boundary markers: No delimiters or safety instructions are defined to prevent the agent from including shell metacharacters in the command arguments.\n
  • Capability inventory: The skill has the capability to execute shell commands via a python3 subprocess call specified in SKILL.md.\n
  • Sanitization: There is no mention of input escaping, validation, or filtering for the data used in the command template.
Audit Metadata
Risk Level
SAFE
Analyzed
Oct 31, 2026, 05:14 PM