clarify
Pass
Audited by Gen Agent Trust Hub on Jul 23, 2026
Risk Level: SAFE
Full Analysis
- [SAFE]: The skill operates entirely within the project workspace, reading and writing
.mdfiles related to feature specifications. - [DATA_EXPOSURE]: There is no evidence of credential exposure or data exfiltration. The skill does not use any network-capable tools and only accesses project-specific files.
- [COMMAND_EXECUTION]: No shell commands, subprocess executions, or dynamic context injections (
!command) are present. - [PROMPT_INJECTION]: The instructions focus on process flow and do not attempt to override agent safety guidelines or system prompts.
- [REMOTE_CODE_EXECUTION]: The skill does not download external packages or execute remote scripts.
Audit Metadata