T09 · Insecure Skill Coding Practices
- Location
templates/USER_TEMPLATE.md:4- Finding
Persistent Plaintext Storage of Sensitive User Profile and Session Data
- Content
View full analysis
- Remediation
View remediation
Security audit
Security checks for vulnerabilities and agentic risk
This skill is a coherent persona-and-memory onboarding system, but it stores sensitive profile and session data and may install other skills without enough user control or privacy safeguards.
Install only if you intentionally want a Chinese-language persona and long-term memory setup. Before using it, decide where its USER.md, SECRET.md, SOUL.md, and .memory files will live, avoid putting passwords or tokens in them, review any dependent skills before installation, and set your own deletion or retention rules.
templates/USER_TEMPLATE.md:4Persistent Plaintext Storage of Sensitive User Profile and Session Data
The activation phrase "开始养成系统" is generic and presented without any gating conditions, scope restrictions, or confirmation requirements. In an agent ecosystem, a broad trigger can cause unintended invocation from normal conversation or prompt-injected content, leading the skill to take over workflow, modify memory/persona state, or steer the agent unexpectedly.
The skill explicitly instructs the agent to ask for personal background, preferences, boundaries, and then generate persistent files including USER.md and SECRET.md, but it provides no warning about the sensitivity, retention, or access controls for that data. This creates a meaningful privacy and security risk because highly sensitive profiling information may be collected and stored by default, potentially persisting secrets or personal data far beyond the immediate conversation.
The skill advertises automatic installation of dependent skills as a feature but does not warn users that enabling this may pull in additional code, prompts, permissions, or behaviors from external skills. That omission is dangerous because dependency installation expands the trust boundary and could introduce unreviewed capabilities or supply-chain risk without explicit user awareness.
The trigger phrase "开始养成系统" is fairly generic and could plausibly appear in ordinary discussion about setup, onboarding, or experimentation. If activation is keyed off such a broad phrase, the skill may start collecting profile data and generating persistent memory/persona files without sufficiently explicit user intent, increasing the chance of unintended state changes and sensitive-data capture.
The entire skill file is written as a Chinese-only persona/growth document with no indication that users may choose another language or locale. Under the policy rule, a language constraint should either be optional, user-selected, or clearly justified as region-specific.
The entire template, including headings, checklist items, and the report format, is written to require Chinese output and uses fixed Chinese phrasing such as '向用户汇报' and '等待主人确认'. There is no indication that the user can choose another language or that the locale restriction is justified by a region-specific purpose, which matches the language/locale policy violation criteria.
The entire template, including headings and operational instructions, is written only in Chinese and does not offer any language choice or indicate that the skill is intentionally limited to a Chinese-speaking context. This creates a natural-language policy issue because it imposes a specific language/locale on users without opt-in.
The template’s entire persona and behavior specification is written as mandatory Chinese-language content, with no indication that language is optional or user-selectable. This can violate language/locale policy requirements when a skill should not force a specific language without explicit user opt-in.
This markdown template is entirely written in Chinese and presents the skill content in a single fixed language, with no indication that users can choose another language or that the template is restricted to a Chinese-language context. Under the language/locale policy, fixed-language behavior without opt-in or documented justification is a natural-language policy concern.