Security audit
Security checks for vulnerabilities and agentic risk
This skill is a Chinese-language amis UI schema helper with templates and a local validator; the reviewed risky-looking items are disclosed and aligned with that purpose.
Installers should treat this as a Chinese-language amis template and validation helper. Before using generated schemas against real backends, review API paths, POST/PUT/DELETE actions, login persistence, and any form data submission behavior.
Skill reads from agent configuration directories (.claude/, .codex/, .gemini/). These directories may contain API keys, personal settings, and other credentials that the skill has no legitimate need to access.
ls ~/.claude/skills/amis-json-skills/templates/
cat ~/.claude/skills/amis-json-skills/templates/basic-crud.json
#### 方式三：对话中触发
The declared description presents the skill as a schema-generation expert that produces AMIS configurations based on business intent and supports advanced interactive behaviors. However, the supplied code chunk is a validation/analysis utility: it contains regex-based pattern libraries for identifying common schema mistakes and returning fixes or suggestions. Its main purpose is linting or reviewing existing schema-like content, not generating new AMIS JSON schemas. While some suggestions touch on mobile adaptation and component configuration, these are narrow support checks and do not match the declared primary capability of end-to-end schema generation with complex interactions.
The declared description emphasizes schema generation expertise: understanding requirements and producing correct, interactive amis configurations. The supplied code does not generate schemas at all. Its primary function is validation of an already-supplied schema, with rule-based checks for page, form, CRUD, dialog, button, wizard, API, and event-action fields, then returning diagnostics. While this is related to amis, it is a materially different purpose from the declared generator role, so this is a clear description-behavior mismatch.
The activation guidance does not define when the skill should load versus when the agent should simply answer normally. In an agent environment, that ambiguity can cause opportunistic or accidental skill activation from common product-related requests, expanding the skill's reach beyond what the user explicitly asked for.
The activation guidance does not define when the skill should load versus when the agent should simply answer normally. In an agent environment, that ambiguity can cause opportunistic or accidental skill activation from common product-related requests, expanding the skill's reach beyond what the user explicitly asked for.
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
# 查看可用模板
ls ~/.claude/skills/amis-json-skills/templates/
复制模板到你的项目：
Skill enumerates or reads other installed skills. Access to other skills' SKILL.md files or the skills directory reveals prompt instructions, capabilities, and secrets that should be invisible to peer skills.
# 查看可用模板
ls ~/.claude/skills/amis-json-skills/templates/
复制模板到你的项目：
The file’s natural-language descriptions and user-facing messages are written entirely in Chinese, starting with the header comment and continuing through the exported pattern messages and suggestions. This can violate a language/locale policy when a skill forces a specific language without user opt-in or documented justification.
This JavaScript file contains natural-language comments and validation messages entirely in Chinese, including the file description and many returned warning/error strings. Because the skill does not indicate that it is intentionally limited to Chinese-speaking users or provide any opt-in language selection, it conflicts with the policy against forcing a specific language without user opt-in.
The natural-language strings in the template header and UI labels are entirely in Chinese, which indicates the skill is designed to present a fixed language experience. There is no indication anywhere in the file that users can opt into this locale or select an alternative language, which can violate language-choice policy for generally applicable skills.
This JSON template contains multiple user-facing strings in Chinese, including the title, card headers, and chart labels, but provides no indication that the language is configurable or user-selected. Under the policy rule for natural-language violations, forcing a specific language without opt-in is a locale policy issue.
This JSON template contains user-facing labels and descriptions entirely in Chinese, such as the title, button label, and field labels. Under the policy rule for natural-language violations, forcing a specific language without opt-in or documented regional justification is a locale-policy issue.
The file header describes the template as a '带过滤搜索的CRUD模板' for '需要条件搜索的数据列表', which frames it primarily as a filtered search/list view. However, the configuration includes write operations for creating, editing, and deleting records via POST, PUT, and DELETE APIs, expanding beyond a search/list template into full data mutation behavior.
This JSON manifest defines a login UI with a checkbox labeled '记住我' and text indicating the user can stay logged in for 30 days. For a markdown/manifest-style skill definition that affects authentication persistence and potentially device privacy, there is no accompanying warning or disclosure about the security implications of using this option on shared devices.
This manifest-like JSON defines a multi-step form that collects name and email and posts the data to /api/submit, but the template contains no visible notice, confirmation text, or warning about transmitting user data. Because the file itself describes behavior affecting user privacy, the lack of disclosure is a semantic safety issue under the missing user warnings rule.
This JSON template uses Chinese-only natural-language labels and placeholders throughout the user interface, such as the title, field labels, and button text. Because the file does not offer a language choice or document that the skill is intentionally limited to a Chinese-speaking or region-specific context, it may violate language/locale policy requirements.
The example '创建一个订单管理页面，需要有搜索筛选功能' is another broad natural-language trigger with no scope limits. Although lower risk than the earlier section, it reinforces a pattern where ordinary requests may silently activate the skill.
The Antigravity runtime example combines explicit mention syntax with a common request phrase but does not state whether the mention is mandatory. That inconsistency can confuse implementers and increase the chance of accidental trigger behavior in MCP-integrated environments.
This markdown file documents supported methods including DELETE and provides a form submission example that posts data and redirects, but it does not include any warning that generated schemas may modify backend data or trigger irreversible actions. For markdown files, the skill description should disclose behaviors that could affect user data or system integrity.
This JSON template uses Chinese titles, labels, placeholders, and confirmation text throughout, which imposes a specific language on users without any visible opt-in or alternative locale handling. The policy for natural-language issues applies to all file types, including config/templates like this one.
Multiple visible strings, including the title, labels, and placeholders, are fixed in Chinese. This can violate language/locale policy when a skill forces a specific language without explicit user opt-in or documentation that the skill is intentionally region-specific.
No suspicious patterns detected.