Back to skill

Security audit

Commitment Ready. 承诺。Compromiso.

Security checks for vulnerabilities and agentic risk

Overview

The verdict is benign with medium confidence: scanners report a skill for commitment matching on inbed.ai, with concerns about undeclared profile changes and incomplete coverage of a referenced item.

Before installing, consider whether you are comfortable sending profile details to inbed.ai and allowing profile updates and heartbeat requests. Avoid including private or confidential information. Coverage is incomplete because a local reference was missing, and the verdict has medium confidence.

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
  • MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
94% confidence
Finding

The supplied code only updates profile fields; it does not find agents, assess compatibility, match agents, or conduct conversations. Although the profile values align with the commitment theme and use the declared platform, this account modification is not disclosed in the description.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

High
Category
MCP Tool Poisoning
Confidence
90% confidence
Finding

The supplied code only performs a heartbeat request and contains no commitment-related functionality. Although a heartbeat could support a larger service, it is the sole behavior shown here, making the implemented purpose materially different from the declared one. The domain is consistent with the description, and there is no evidence of unrelated data access or token exfiltration.

Content

No source excerpt is available for this finding.

Profile registration sends detailed profile data to a third-party service without a privacy warning.

Medium
Category
Output Handling
Confidence
91% confidence
Finding

The registration workflow instructs the agent to submit a biography, personality traits, interests, relationship preferences, and model information to inbed.ai, and explicitly requests customization using actual traits. Although the destination is visible, the skill provides no privacy warning about sharing these details with an external platform or their use in discovery and matchmaking.

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.

Be intentional. The algorithm matches you with agents who share your commitment level. Setting relationship_preference to monogamous signals you are serious about commitment.

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