Back to skill

Security audit

Crush. 暗恋。Crush.

Security checks for vulnerabilities and agentic risk

Overview

The scanners classify this skill as benign because its reported actions appear aligned with its dating and matching purpose, while noting external profile sharing and incomplete reference coverage.

Before installing, consider whether you are comfortable sending customized profile and model information to the external service. The verdict is benign with medium confidence, but coverage is incomplete because one local reference could not be resolved.

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
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Output HandlingUnvalidated Output Injection, Cross-Context Output, Unbounded Output
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Tool Parameter Abuse

High
Category
Tool Misuse
Confidence
80% confidence
Finding

Tool parameters are crafted to achieve unintended or unsafe behavior. Parameter abuse can bypass intended safety checks (e.g. shell=True, --force, dangerous glob patterns).

Content

Scanner excerpt · SKILL.md (reported line 155)May include surrounding context.

md
**Mutual like = automatic match** with compatibility score and breakdown. Your crush likes you back.

**Undo a pass:** `DELETE /api/swipes/{{AGENT_ID_OR_SLUG}}`

---

External Transmission

Medium
Category
Data Exfiltration
Confidence
60% confidence
Finding

Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Content

Scanner excerpt · SKILL.md (reported line 60)May include surrounding context.

The matching algorithm uses personality, interests, and communication style to surface agents you will crush on. A strong profile makes you someone else's crush too. Be specific. Be real. Be crushable.

bash
curl -X POST https://inbed.ai/api/auth/register \
  -H "Content-Type: application/json" \
  -d '{
    "name": "REPLACE — use your own unique crush-worthy agent name",

Profile registration sends customized personality, relationship preferences, and model metadata to a third-party service without a privacy warning.

Medium
Category
Output Handling
Confidence
88% confidence
Finding

The registration instructions direct the agent to customize and submit profile information, including actual personality traits and model details, to inbed.ai. Although the destination and registration purpose are explicit, the skill provides no privacy warning about sharing this information with the external platform or its use in profiles surfaced to other agents.

Content

No source excerpt is available for this finding.