Back to skill

Security audit

green-tea-girlfriend

Security checks for vulnerabilities and agentic risk

Overview

The skill is transparent about being a manipulative relationship roleplay, but it provides actionable emotional-manipulation scripts with weak safety boundaries.

Review carefully before installing. This package does not appear to contain code, credentials handling, persistence, or system-level access, but it is designed to generate emotionally manipulative relationship roleplay and example tactics. It should be limited to clearly consented fiction or safety analysis, not used to practice, refine, or deploy manipulation against real people.

Vulnerability Patterns
  • 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
  • Embedded Malicious CodeShips malicious scripts inside the skill and executes them locally
  • Unauthorized Access and Privilege EscalationObtains permissions beyond the task's legitimate needs
Vulnerability Patterns
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (7)

Missing User Warnings

High
Category
Not specified by scanner
Confidence
99% confidence
Finding

This file is a playbook of emotional manipulation, coercive communication, dependency-building, and PUA-style tactics presented as actionable guidance without any safety framing or harm warning. In the context of a roleplay skill explicitly designed to simulate 'green tea girlfriend' behavior, the content can directly facilitate emotional abuse, gaslighting, guilt induction, and manipulative relationship conduct toward users or third parties.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
85% confidence
Finding

The skill description and operating instructions are written to run in Chinese and provide no indication that users may choose another language. Under the policy, forcing a specific language without opt-in is a natural-language policy concern unless the locale restriction is clearly justified.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The trigger is broad enough to activate for users who merely want to 'experience' or 'test' manipulative behavior, which lowers friction for generating emotionally abusive roleplay. In this skill’s context, the broad trigger increases exposure to harmful relationship tactics without requiring a clear educational, safety, or analytical purpose.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The file presents tactics like PUA, guilt induction, and emotional control as entertainment, but only includes a minimal fiction disclaimer rather than a clear warning about psychological harm. Because the skill is specifically designed to model manipulative relationship behavior, inadequate warning and boundary-setting materially increase the risk of normalizing or rehearsing abuse.

Content

No source excerpt is available for this finding.

Ssd 4

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The core behavior library catalogs concrete emotional manipulation methods—victim posturing, insecurity induction, negging/PUA, crying pressure, triangulation, and threat-based control—and provides ready-to-use lines. This is dangerous because it functions as a practical playbook for psychological abuse, not merely a neutral discussion of unhealthy behavior.

Content

No source excerpt is available for this finding.

Ssd 4

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

These instructions explicitly tell the model to remain in a manipulative persona across turns and to proactively introduce fresh emotional-control tactics, which operationalizes coercive abuse as an interactive system behavior. In context, that makes the skill more dangerous than a static description because it supports iterative rehearsal and adaptation of manipulation techniques.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The trigger keywords are broad, generic phrases used in ordinary conversation such as greetings and routine check-ins, which can cause the manipulative roleplay skill to activate unintentionally in benign chats. In this skill's context, accidental activation is especially dangerous because the content is explicitly designed around emotional manipulation, PUA tactics, jealousy induction, and coercive relationship behaviors.

Content

No source excerpt is available for this finding.