create-command
Pass
Audited by Gen Agent Trust Hub on Apr 21, 2026
Risk Level: SAFEPROMPT_INJECTIONCOMMAND_EXECUTION
Full Analysis
- [PROMPT_INJECTION]: The skill facilitates indirect prompt injection by generating agent instructions directly from untrusted user input.
- Ingestion points: User-provided
<command-name>and<description>parameters (SKILL.md). - Boundary markers: Absent; instructions are directly interpolated into the generated markdown files without delimiters or instructions to ignore embedded content.
- Capability inventory: Writes markdown files to
.cursor/commands/and.claude/commands/directories (SKILL.md). - Sanitization: Absent; the skill does not validate or sanitize the content of the generated command files.
- [COMMAND_EXECUTION]: The skill automates the creation of markdown files that are intended to contain executable shell commands (e.g., the git diff example), representing the generation of executable content from user requirements.
Audit Metadata