Security audit
Security checks for vulnerabilities and agentic risk
The package is a disclosed local agent-governance plugin, but users should review its hook/conversation access and note that the published artifact appears to lack its declared runtime entrypoint.
Install only if you want a local behavior-governance layer that can observe agent conversations/tool activity and persist evidence for future guidance. Check the language setting, keep telemetry disabled unless you explicitly want it, and disable conversation access if that scope is too broad. Because the declared dist runtime files were absent in this artifact, verify the installed package contents before relying on it.
The manifest description at L03 scopes the skill to health checks, pain recording, viewing task/run/candidate/artifact data, candidate intake, flow policy viewing, and pain-to-principle traceability. However, the file also documents activation listing and deactivation, which are operational state-changing capabilities not mentioned in that stated scope.
The natural-language description is written as a directive and includes no option for another language or locale. Under the policy, forcing a specific language without user opt-in is a reportable language/locale violation.
The skill is written to operate in Chinese throughout and does not offer any language selection or fallback, which can exclude or mislead users who interact in another language. In a command-guidance skill, this raises safety and usability risk because users may misunderstand operational or security-relevant instructions, leading to incorrect command execution or failure to understand permission boundaries.
The natural-language content of the skill, including its description and operating instructions, forces a specific language/locale for users. The policy allows language constraints only when the skill offers a choice or clearly documents a justified region-specific limitation, which is not present here.
The skill is written entirely in Chinese and directs structured output in Chinese without indicating that language is configurable or required by the operating environment. This can exclude users, reduce operator comprehension in mixed-language deployments, and increase the chance that reviewers miss unsafe implementation details or misunderstand execution/reporting requirements.
No suspicious patterns detected.
