create-cli
Warn
Audited by Gen Agent Trust Hub on Mar 10, 2026
Risk Level: MEDIUMCOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: In Phase 2 (Audit), the skill explicitly directs the agent to "Run --help via Bash to capture actual behavior." Since the
<cli>identifier is supplied by the user, this provides a direct path for arbitrary command execution if the input contains shell metacharacters, command substitution, or pipes. - [PROMPT_INJECTION]: The skill has an indirect prompt injection surface as it ingests and processes untrusted data from user-specified file paths using auditing tools.
- Ingestion points: User-provided CLI source locations and files analyzed via
Glob,Grep, andReadin the Audit phase. - Boundary markers: Absent; there are no instructions to the agent to treat content from the audited files as data rather than instructions.
- Capability inventory: The skill utilizes
Bash,Write,Read,Glob, andGrep, allowing for significant system interaction if an injection attack succeeds. - Sanitization: Absent; the skill does not perform any validation or sanitization on the file contents or paths before processing.
Audit Metadata