cbi-script
Pass
Audited by Gen Agent Trust Hub on May 5, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill utilizes the vendor-provided
cbicommand-line interface to perform operations such as creating, retrieving, and saving project scripts (e.g.,cbi project script-save). This involves the execution of shell commands based on user input and project state. - [PROMPT_INJECTION]: The skill exhibits an indirect prompt injection surface because it retrieves script content from an external system and may use that content in subsequent processing steps.
- Ingestion points: Script content fetched from the project database via
cbi project script-get. - Boundary markers: The instructions do not specify any delimiters or warnings to ignore instructions embedded within the retrieved script content.
- Capability inventory: Shell command execution via the
cbiCLI toolset (create, get, save). - Sanitization: There is no evidence of sanitization or safety validation applied to the content fetched before it is processed or re-saved.
Audit Metadata