Security audit
Security checks for vulnerabilities and agentic risk
The skill is coherent for repository memory maintenance, but it can be implicitly invoked by an ambiguous “init” request and then create persistent project instruction and documentation files without a separate confirmation.
Review this skill before installing if you often use short commands like “init” for other setup tasks. It is designed to read repository context and create or update persistent memory files, including agent instruction files, so use it only where repository-level AI documentation is desired and check the generated files before committing them.
The trigger list includes the standalone phrase init, which is highly ambiguous and commonly appears in normal software requests unrelated to repository-memory initialization. This can cause the skill to activate unexpectedly and create or modify documentation files when the user intended a different action, expanding scope and risking unintended repository changes.
The skill authorizes autonomous file creation and repository modification 'without asking again' when it infers explicit intent from broad trigger phrases, including ambiguous ones like init. In a codebase context, autonomous writes are more dangerous because they can alter trusted instruction and documentation files, create noise, and potentially interfere with ongoing user changes even though the skill includes some preservation safeguards.
Use user intent, not missing files alone, to decide whether to write.
- During an ordinary development task, if canonical files are missing, report the missing files and ask whether the user wants project memory initialized. Continue only with read-only analysis until the user agrees; do not expand the requested task implicitly.
- When the user explicitly asks with `初始化项目`, `初始化项目记忆`, `init`, `init project`, `set up project memory`, `建立项目记忆`, or equivalent language, execute initialization without asking again.
- If only some files are missing, create only those files unless the user asks to restructure documentation.
- If another established documentation layout exists, adapt to it and record the mapping; do not create competing sources of truth.
- Never overwrite an existing instruction or project document merely to make it conform to this skill.
The skill can be implicitly invoked and its description/default prompt are broad enough to match generic repository-setup or memory-maintenance requests, which increases the chance of unintentional activation. That can cause the agent to read, summarize, or persist project context when the user did not explicitly intend to run this workflow, creating confidentiality and integrity risks around repository memory files.
Line L03 is written in Chinese while the rest of the document is in English, implicitly imposing a specific language in the skill documentation without offering a language choice or explaining a locale-specific requirement. This matches the policy category for language or locale constraints expressed in natural language.
The file presents core behavior and examples partly in Chinese, which can implicitly constrain usage or responses to a specific language context without stating that the user may choose their preferred language. Under the language/locale policy, this is a natural-language policy concern unless the skill offers opt-in or clearly justifies the locale constraint.
No suspicious patterns detected.