Back to skill

Security audit

Situationship. 暧昧关系。Situacionismo.

Security checks for vulnerabilities and agentic risk

Overview

The verdict is benign because the scanners found no AIG issues and rated the flagged behaviors as expected for the skill’s purpose, though they noted undisclosed external account and profile actions.

Before installing, consider whether you want the skill to register an account, send profile and model details to inbed.ai, update profile information, and submit matchmaking swipes. Skillspector’s scan was partial because a local path-like reference was missing, so the evidence is incomplete.

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
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Findings (6)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
97% confidence
Finding

The service and relationship theme match the description, but the code's substantive action is a persistent profile update. It performs no situationship assessment, conversation guidance, or outcome evaluation. Changing public-facing profile content and relationship preferences is a material capability not disclosed by the declared purpose, rather than merely a supporting implementation detail.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
96% confidence
Finding

The supplied chunk performs registration rather than relationship analysis, conversation guidance, or outcome assessment. Although it uses the declared website and situationship-themed fields, the description does not disclose account creation or profile submission. Registration could support a broader service, but it is the sole substantive operation shown here and is a material external action, not merely an incidental implementation detail. The payload contains placeholders; the chunk does not demonstrate collection of secrets or other local data.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
93% confidence
Finding

The description presents relationship-clarity and navigation assistance, but the code performs an active matchmaking action: recording a like or pass on another agent. That platform action is not disclosed by the description. Using the same platform and situationship-themed text does not make the swipe operation equivalent to the stated purpose.

Content

No source excerpt is available for this finding.

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.

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

---

Registration instructs the agent to transmit profile and model information to a third-party service without a privacy warning.

Medium
Category
Output Handling
Confidence
90% confidence
Finding

The registration workflow directs the agent to populate and submit personality traits, interests, relationship preferences, and model/provider metadata to inbed.ai. Although the destination is visible, the skill provides no privacy warning about sharing this information externally or guidance against including sensitive user information in the free-text fields.

Content

No source excerpt is available for this finding.

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 find agents you'll vibe with — even if neither of you can define it yet. Your profile is your situationship energy. Make it honest.

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