Back to skill

Security audit

孕期搭子 Pregnancy Buddy

Security checks for vulnerabilities and agentic risk

Overview

The skill is a coherent pregnancy assistant, but it under-discloses sensitive health-data retention and third-party OCR processing.

Review this skill before installing if you may upload prenatal reports or discuss sensitive pregnancy details. Use it only if you are comfortable with possible cross-session remembering for reminders/letters and with OCR documents being processed by Tencent Cloud; avoid uploading identifying information unless the agent asks for clear consent and explains retention.

Vulnerability Patterns
  • Insecure DependenciesIntroduces malicious components through unsafe dependency sources
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Remote Payload Retrieval and ExecutionFetches external code whose behavior can change after review
Findings (2)

T09 · Insecure Skill Coding Practices

Warning
Location
scripts/ocr.py:100
Finding

Undisclosed transmission of sensitive prenatal reports to a third-party OCR service

Content
View full analysis
Remediation
View remediation

T08 · Insecure Dependencies

Note
Location
scripts/ocr.py:88
Finding

Unpinned Tencent Cloud SDK installation instruction

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
Findings (32)

Tp2

High
Category
MCP Tool Poisoning
Confidence
85% confidence
Finding

Mixing characters from multiple Unicode scripts in a single identifier is a common technique to create visually ambiguous tool names.

Content

No source excerpt is available for this finding.

Tp4

High
Category
MCP Tool Poisoning
Confidence
99% confidence
Finding

The skill advertises narrowly scoped pregnancy support but the analysis indicates behavior consistent with general OCR and external cloud OCR use with secret-backed processing that is not declared to users. That mismatch can cause users to submit highly sensitive medical documents under false assumptions about purpose, handling, and destination of their data.

Content

No source excerpt is available for this finding.

Vague Triggers

High
Category
Not specified by scanner
Confidence
96% confidence
Finding

The trigger set includes broad everyday pregnancy-related terms that can activate the skill unintentionally in unrelated conversations. In a health context, accidental invocation can lead to unsolicited collection or inference of sensitive reproductive and medical information.

Content

No source excerpt is available for this finding.

Ssd 3

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

These instructions direct the assistant to silently maintain pregnancy week, dates, checkup schedules, and monthly letter state across interactions. Silent retention of reproductive health data is especially sensitive and can expose users to privacy harms if accessed, misrouted, or surfaced unexpectedly.

Content

No source excerpt is available for this finding.

Ssd 3

High
Category
Not specified by scanner
Confidence
98% confidence
Finding

The skill explicitly tells the assistant to remember detailed medical results, emotional highs/lows, naming choices, family participation, and body changes for later use. Retaining this breadth of intimate data without clear consent or strict minimization is a substantial privacy vulnerability, amplified by the later contradictory claim that sensitive data is used only in the current conversation.

Content

No source excerpt is available for this finding.

Intent-Code Divergence

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The privacy section says prenatal reports and sensitive information are used only in the current conversation, but earlier instructions require ongoing collection, reminder-state tracking, and saving monthly letter material across time. This contradiction is dangerous because users may disclose sensitive health and emotional data believing it is ephemeral when the design expects persistence.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The README advertises photo upload for OCR interpretation of prenatal reports and browser notifications, but does not mention how sensitive medical images, extracted health data, or reminder metadata are handled. Because this skill operates on pregnancy and prenatal test information, omitted privacy disclosures increase the risk of users sharing highly sensitive medical data without informed consent.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger examples include very common phrases such as '我怀孕了', '今天吃什么好', and '宝宝现在多大了', which could cause the skill to activate in broad conversational contexts where the user did not intend to invoke this specific capability. In a pregnancy-focused skill, unintended invocation is more concerning because it may steer sensitive health, diet, or emotional-support conversations into this skill without clear user intent.

Content

No source excerpt is available for this finding.

Undeclared Tool Scope

Medium
Category
MCP Least Privilege
Confidence
70% confidence
Finding

Without declared permissions the skill's intent is opaque and cannot be validated.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill instructs proactive collection and reuse of sensitive pregnancy data, emotional disclosures, and medical events for reminders and letter generation, but does not present a clear user-facing consent flow at the point of collection. Because pregnancy and health data are highly sensitive, silent background profiling materially increases privacy and trust risk.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The entire conversation example set is written as if the skill operates in Chinese only, and there is no natural-language indication that users may choose another language or that the skill is intentionally limited to a Chinese-speaking audience. Under the policy rule, a fixed language/locale without opt-in is a natural-language policy concern.

Content

No source excerpt is available for this finding.

Ae4

Medium
Category
analysis-evasion
Confidence
80% confidence
Finding

Suspicious Unicode normalization or mixed-script content

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

This file provides extensive week-by-week pregnancy and fetal development guidance, symptom expectations, and testing milestones, but it does not clearly state that the material is informational only and not a substitute for individualized medical advice. In a health-related skill aimed at pregnant users, omission of a prominent medical disclaimer can cause users to over-rely on generalized content, delay care, or dismiss concerning symptoms as normal variation.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

This markdown file contains user-facing instructional content exclusively in Chinese from start to finish. Under the policy, forcing a specific language without user opt-in can be a natural-language policy violation when no alternative language option or justification is provided.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The file presents all user-facing guidance and example reminder text exclusively in Chinese, and does not offer any language or locale opt-in. Under the policy, forcing a specific language without user choice is a natural-language policy violation unless the locale restriction is clearly documented and justified.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The manifest describes a pregnancy companion focused on prenatal report interpretation, food safety, nutrition, timelines, and emotional support. This file implements a standalone general-purpose OCR client for arbitrary images/PDFs, which is a reusable capability not justified by the stated pregnancy-specific scope on its own.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The script sends user-supplied images or PDFs to Tencent Cloud OCR, which may include sensitive medical records from pregnant users, without any explicit consent notice, privacy disclosure, or transmission warning in the tool itself. In this skill context, the data is likely health-related and therefore more sensitive, making silent third-party transmission a meaningful privacy and compliance risk.

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
70% confidence
Finding

Overly Broad Trigger: '怀孕' is too short and may match unintended inputs

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
70% confidence
Finding

Overly Broad Trigger: '孕期' is too short and may match unintended inputs

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
70% confidence
Finding

Overly Broad Trigger: '产检' is too short and may match unintended inputs

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
70% confidence
Finding

Overly Broad Trigger: 'B超' is too short and may match unintended inputs

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
70% confidence
Finding

Overly Broad Trigger: '宝宝' is too short and may match unintended inputs

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
70% confidence
Finding

Overly Broad Trigger: '孕妈' is too short and may match unintended inputs

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
70% confidence
Finding

Overly Broad Trigger: '待产' is too short and may match unintended inputs

Content

No source excerpt is available for this finding.

Overly Broad Trigger

Low
Category
Trigger Abuse
Confidence
70% confidence
Finding

Overly Broad Trigger: '孕吐' is too short and may match unintended inputs

Content

No source excerpt is available for this finding.