openspec-ff-change
Pass
Audited by Gen Agent Trust Hub on Mar 7, 2026
Risk Level: SAFECOMMAND_EXECUTIONPROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes local shell commands using the
openspecCLI to manage change artifacts. Commands includeopenspec new,openspec status, andopenspec instructions. These commands utilize arguments dynamically generated from user input. - [PROMPT_INJECTION]: The skill is susceptible to indirect prompt injection and potential command injection through user-provided names or descriptions.
- Ingestion points: Step 1 collects a change name or description directly from the user.
- Boundary markers: No explicit shell escaping or strict boundary markers are defined for the command line arguments in the instructions.
- Capability inventory: The skill has the capability to create directories and files via the
openspecCLI and theTodoWritetool. - Sanitization: The instructions direct the agent to 'derive a kebab-case name' from the user input. While this serves as a sanitization step, it relies on the AI agent's adherence to the instruction to prevent malicious characters from reaching the shell.
Audit Metadata