openspec-new-change
Pass
Audited by Gen Agent Trust Hub on Sep 18, 2026
Risk Level: SAFECOMMAND_EXECUTIONINDIRECT_PROMPT_INJECTION
Full Analysis
- [COMMAND_EXECUTION]: The skill executes openspec CLI commands using arguments derived from user input (e.g., openspec new change ""). These commands are used to scaffold project directories and retrieve status information.
- [INDIRECT_PROMPT_INJECTION]: The skill processes user-provided descriptions to generate command arguments. It includes guardrails to mitigate risks associated with processing untrusted input.
- Ingestion points: User input via AskUserQuestion for the change name or description in SKILL.md.
- Boundary markers: The skill instructions mandate that the agent validates the derived name follows a specific kebab-case format.
- Capability inventory: Executes local openspec commands (new change, status, instructions).
- Sanitization: Instructs the agent to verify the name format and uses double-quoting around the variable in the shell execution context.
Audit Metadata