jimeng-opencli-text2image
Pass
Audited by Gen Agent Trust Hub on Jul 28, 2026
Risk Level: SAFENO_CODECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [NO_CODE]: The skill consists exclusively of markdown instructions and does not include any scripts, binaries, or configuration files.
- [COMMAND_EXECUTION]: The instructions direct the agent to execute shell commands using the opencli utility (e.g., opencli jimeng generate).
- [PROMPT_INJECTION]: The skill uses specific instructional language to constrain the agent's behavior and tool selection, such as forbidding the use of the dreamina CLI or alternative execution-class skills.
- [COMMAND_EXECUTION]: A vulnerability surface for potential command injection was identified in how prompts are handled.
- Ingestion points: The arguments in the generate and generate-image2image commands described in SKILL.md.
- Boundary markers: While documentation examples wrap prompts in double quotes, there are no explicit instructions for the agent to sanitize or escape shell metacharacters (e.g., backticks, semicolons) within the input.
- Capability inventory: The skill facilitates direct invocation of shell commands via the opencli utility.
- Sanitization: No sanitization, filtering, or validation logic is specified for the external prompt content before it is passed to the shell.
Audit Metadata