Security audit
Security checks for vulnerabilities and agentic risk
This media-generation skill is mostly purpose-aligned, but it should be reviewed because it encourages users to paste a reusable API key into chat and persist it for later use.
Review this skill before installing. Use a secure configuration path for RAISE_AI_API_KEY instead of pasting the key into chat, rotate the key if it is exposed, and avoid submitting sensitive media or sharing signed output links unless you are comfortable with the RaiseAI service and temporary-link access model.
The skill forces activation for a very broad class of image/video/script-related requests, including cases where the user did not explicitly request this provider or tool. Overbroad routing can cause unintended external data disclosure to a third-party service, reduce user choice, and trigger actions with side effects or cost when a narrower or safer response would suffice.
The skill instructs users to disclose their API key directly to the agent for automatic storage, which is a classic secret-handling anti-pattern. In the context of an agent skill, this is more dangerous because the agent may log, persist, summarize, or route the secret across multiple components, increasing the chance of credential exposure and unauthorized API use.
The document repeats that users can configure the skill by telling the agent their API key and that it will be persisted as an environment variable. Repetition normalizes unsafe secret sharing and increases the likelihood that users will expose reusable credentials in chat, where they may be retained or accessible beyond the intended scope.
The skill explicitly tells users they can paste their RaiseAI API key into the conversation for automatic configuration, which encourages secret handling through normal chat channels. This increases the risk of credential exposure in logs, transcripts, downstream tools, or accidental model reuse, especially because the skill does not present a strong warning or safer alternative at the point of collection.
The skill not only invites a user to provide an API key in chat, but also states it will be stored in an environment variable for later use. Accepting and persisting user-declared credentials from conversational input creates a direct secret-ingestion path that can be abused, mishandled, or retained beyond user expectations, enabling unauthorized API usage if exposed.
The documentation explicitly tells users to paste their RaiseAI API key into the agent so it can auto-configure and persist the secret. This encourages disclosure of sensitive credentials through a conversational channel, which may be logged, retained, exposed to other tools, or mishandled by downstream components.
The documentation instructs users to provide publicly accessible image and video URLs to a third-party API but does not warn that making media publicly reachable can expose sensitive or proprietary content. In this skill context, users are specifically encouraged to upload/analyze personal media, which increases the likelihood of privacy leakage, unintended disclosure, and persistence of accessible files beyond the task duration.
The file is written entirely in Chinese and does not provide any language choice, fallback, or justification for restricting instructions to that locale. Under the policy, forcing a specific language without user opt-in can be considered a natural-language policy violation.
The documentation instructs the agent to return full signed temporary URLs, including access parameters, to users without any warning that these links act as bearer-style access tokens during their validity period. If such links are copied into chats, logs, analytics, or shared with unintended parties, anyone holding them may access the generated media until expiration, creating a privacy and unintended disclosure risk.
SQP-3 applies to all file types and covers language or locale policy violations. This file presents all operational guidance exclusively in Chinese, with no opt-in, alternative language, or stated region-specific justification, which may force a specific language on users.
All natural-language instructions and examples in the file are presented exclusively in Chinese, with no indication that users may choose another language or that the scope is intentionally limited to a Chinese-language audience. This can constitute a language-policy issue when a skill implicitly forces a specific language without user opt-in.
This markdown file contains only Chinese-language instructions and examples, with no indication that users may choose another language or that the scope is intentionally Chinese-only. Under the policy rule, forcing a specific language without opt-in is a natural-language policy concern.
No suspicious patterns detected.
