Security audit
Security checks for vulnerabilities and agentic risk
This skill is a disclosed TDD workflow helper that can run local development commands and create local Git commits, with no evidence of hidden exfiltration, persistence, or destructive behavior.
Install this only for repositories where you are comfortable letting the agent run local development commands, edit code and plan files, run formatters, and create local Git commits. Review changes before pushing them or using it on sensitive projects.
The skill advertises automatic task execution, code commits, and plan-file updates without an explicit upfront warning that it will modify repository contents and Git history. In a tool with read and exec permissions, silent or insufficiently disclosed state changes are dangerous because users may trigger destructive or hard-to-reverse actions without informed consent.
The scope language is overly broad and includes generic project-management and automation scenarios, which increases the chance the skill will activate in contexts beyond its intended use. Because the skill has exec capability and can modify plans, code, and Git state, over-broad triggering raises the risk of unintended command execution and repository changes from ambiguous user requests.
The skill claims no extra API key is needed because it only uses local Git and file operations, but it also explicitly depends on an external LLM API provided by the agent platform. This can mislead users about trust boundaries, data exposure, and operational dependencies, causing them to run the skill without understanding that repository contents or prompts may be sent to a remote model service.
The license section states the skill is '完全中文化表达，适配中文独立开发者工作流', indicating a fixed language/locale orientation. There is no accompanying language choice, opt-in, or explicit justification that the tool is region-specific, so this is a natural-language locale policy concern.
