Security audit
Security checks for vulnerabilities and agentic risk
This skill is a narrow, disclosed wrapper for fetching a public daily news image from UAPI, with no local code execution, persistence, or hidden data access.
Before installing, be aware that broad phrases like daily digest may activate this skill and make a request to UAPI for a news image. It appears safe for its stated purpose, but users who want strict control over external calls may prefer explicit invocation or narrower triggers.
The trigger list includes broad, everyday phrases such as "daily digest" and "news poster," which can match requests that are not specifically about this UAPI endpoint. This increases the chance of unintended skill invocation, causing the agent to select a narrower external API action than the user intended and potentially route data or workflow decisions incorrectly.
The manifest sets allow_implicit_invocation: true, which permits the skill to be triggered without tightly constrained user intent or explicit invocation boundaries. Even though this skill only targets a read-style GET endpoint for a daily news image, broad implicit activation increases the chance of unintended tool use, parameter misuse, or prompt-driven invocation when the user did not clearly request this external action.
The natural-language description is presented in both Chinese and English, but the document does not state how the assistant should choose the response language or defer to user preference. This can conflict with language/locale policy expectations when a skill effectively imposes a language choice without explicit opt-in.
This markdown file contains only Chinese-language user-facing instructions and descriptions, with no indication that users may choose another language or that the locale restriction is intentional. Under the policy rule, forcing a specific language without opt-in is a natural-language policy concern.
The file is primarily written in Chinese, but the keyword index also prescribes English trigger phrases. This creates an implicit language behavior without clarifying whether the skill should match one language, both, or based on user preference, which can conflict with language/locale policy expectations.
