openspec-verify-change
Pass
Audited by Gen Agent Trust Hub on Apr 2, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: No malicious behavior or security violations were identified. The skill's operations are consistent with its documented purpose.
- [COMMAND_EXECUTION]: The skill executes commands using the
openspecCLI to fetch change lists and statuses. These operations are restricted to the local environment and use variables derived from user-selected change names. - [DATA_EXFILTRATION]: The skill accesses local files such as
tasks.md,design.md, and specification files to compare them against the implementation. It does not contain any network-enabled tools or instructions to transmit this data externally. - [INDIRECT_PROMPT_INJECTION]: The skill ingests data from external project artifacts. It analyzes these files for specific structural elements like checkboxes and requirement headers to build a report. It does not treat the content of these files as instructions to the agent, which effectively mitigates the risk of indirect prompt injection.
Audit Metadata