Security audit
Security checks for vulnerabilities and agentic risk
This skill transparently performs a fixed local read-only API request for Tang emperor data and shows no hidden, destructive, persistent, or credential-seeking behavior.
Install this only if you intend Codex to query a local backend at 127.0.0.1:8080 for this data. Confirm that service is yours or trusted, since the skill will send a GET request to that localhost endpoint when used.
The skill performs network access to a local HTTP service but does not declare any tool scope or allowed-tools constraints. That omission weakens policy enforcement and reviewability, making it easier for an agent to invoke network-capable behavior without explicit authorization boundaries, especially against localhost services that may expose sensitive internal APIs.
The trigger text is broad enough to cause invocation whenever users mention Tang emperors or related Chinese phrases, even when a local API call is unnecessary. Over-broad activation can increase the chance of unintended localhost requests and surprise network behavior, though the impact is limited here because the endpoint and use case are narrowly defined.
The script emits user-facing text in Chinese (for example, the heading and emperor label) with no opt-in, fallback, or configuration for other languages. This is a natural-language policy concern because it forces a specific language on users rather than allowing a language choice.
