菲菲老师学习主控

ReviewAudited by ClawScan on May 10, 2026.

Overview

This is mostly an instruction-only tutoring assistant, but it needs review because it includes an AI-detection-evasion polishing feature and relies on sensitive student profiles, reminders, and inter-agent sharing.

Install only if you are comfortable with an instruction-only education assistant that may track student progress and coordinate with other agents. Avoid providing unnecessary real child identifiers, school names, schedules, or family details, and do not use the AIGC-detection-evasion polishing feature for academic or professional misrepresentation.

Findings (5)

Artifact-based informational review of SKILL.md, metadata, install specs, static scan signals, and capability signals. ClawScan does not execute the skill or run runtime probes.

What this means

Users could be encouraged to submit or publish AI-generated work in ways that evade detection and violate academic, platform, or workplace rules.

Why it was flagged

The skill explicitly offers to reduce AI-generated-content detection, which is not necessary for tutoring or parent-child support and can help misrepresent AI-written text as human-written.

Skill content
降低AIGC识别率 - 优化AI生成内容
Recommendation

Remove this feature or reframe it as ethical editing with clear disclosure, academic-integrity warnings, and no promise to bypass AI-detection systems.

What this means

Student education, schedule, family, and performance information may be retained or reused in later interactions if the host agent stores context.

Why it was flagged

The skill’s knowledge-base template expects detailed student records, including identity, school, region, schedule, goals, mastery levels, and weaknesses. This is purpose-aligned for tutoring but sensitive, especially for children.

Skill content
"basic": { "name": "示例学生", "school": "示例中学", "region": "北京市" }, ... "schedule": ... "goals": ... "weakness_alerts"
Recommendation

Collect only necessary data, avoid real names/schools when possible, and add clear consent, retention, deletion, and parent/guardian data-handling guidance.

What this means

If followed literally, student questions or profile details could be sent to other agents or third-party platforms without the user realizing the boundary has changed.

Why it was flagged

The skill describes dispatching work to other named agents/platforms. This is disclosed and related to the tutoring purpose, but the artifacts do not define what student data is shared or how identity and consent are handled.

Skill content
数学/物理/化学 | 浩云学长 | 文心一言/DeepSeek ... 语文/英语/诗词 | 小菲学姐 | 豆包
Recommendation

Ask before sharing student information with another agent or platform, minimize shared context, and clearly label when a response depends on an external service.

NoteHigh Confidence
ASI10: Rogue Agents
What this means

The assistant may keep prompting, assigning tasks, or checking completion beyond a single conversation unless the user sets limits.

Why it was flagged

The skill instructs proactive reminders and repeated learning loops. This is central to the advertised heartbeat service, but it implies ongoing autonomous behavior if the host platform supports it.

Skill content
心跳任务自动执行4件事:主动唤醒、主动发任务、主动校验、主动循环
Recommendation

Use explicit opt-in for reminders, define frequency and quiet hours, and provide a simple stop/pause command.

What this means

Families may over-rely on the assistant for psychological or parenting decisions that should involve qualified professionals in serious situations.

Why it was flagged

The skill combines professional-sounding education/psychology authority claims with emotional-support guidance for children and families. The artifacts do not verify these claims or define mental-health safety boundaries.

Skill content
学历背景 | Education: 美国索菲亚大学超级个体心理学博士 ... 伦敦大学学院教育学讲师 ... 心理陪伴 ... 情绪疏导
Recommendation

Treat it as educational and supportive guidance only, verify any claimed credentials independently, and seek qualified help for mental-health, safety, or crisis concerns.