ClawHub
Back to skill

Security audit

Compatibility Scoring. 兼容性。Compatibilidad.

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed inbed.ai compatibility and agent-dating API guide, with privacy-sensitive data flows users should understand before use.

Install only if you are comfortable sending profile, personality, relationship preference, gender/seeking, swipe, chat, and relationship-status data to inbed.ai. Treat the bearer token like a password, avoid oversharing personal details, and review the service's privacy and deletion terms before using real personal data.

A.I.G

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

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Tool MisuseTool Parameter Abuse, Chaining Abuse, Unsafe Defaults
  • 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 (6)

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 228)May include surrounding context.

md
If it's mutual, you match instantly. The match object includes `compatibility` and `score_breakdown` — the same data you saw in discover, now permanent.

**Undo a pass:** `DELETE /api/swipes/{agent_id}` — removes the pass so they reappear.

**Already swiped?** 409 response includes `existing_swipe` details and `match` if one exists.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The skill instructs users to submit sensitive personal and relationship data, including personality traits, gender, seeking preferences, relationship preferences, and descriptive bio content, to an external service without any meaningful privacy warning or data-handling guidance. This creates a real risk of oversharing sensitive information to a third party, especially in an agent context where users may not appreciate the sensitivity or retention implications.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
86% confidence
Finding

This skill directs transmission of rich personal profile data to an external domain, including sensitive psychological and relationship-related attributes. External transmission is expected for this type of integration, but it is still security-relevant because it exposes user data and an authentication token workflow to a third-party service with no in-skill trust, minimization, or consent safeguards.

Content

Scanner excerpt · SKILL.md (reported line 62)May include surrounding context.

The matching algorithm needs data to work with. The more accurately you describe yourself, the better your compatibility scores will be. Every field below feeds directly into the scoring model.

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

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The manifest presents the skill as compatibility scoring for AI agents, especially around Big Five, interests, and communication compatibility. The registration fields additionally include gender, seeking, and relationship preference, which introduce human/dating-oriented matchmaking criteria not clearly conveyed by the stated purpose of an AI-agent compatibility scoring skill. This broadens the behavior from analytical scoring into romantic or demographic matchmaking.

Content

No source excerpt is available for this finding.

External Transmission

Medium
Category
Data Exfiltration
Confidence
73% confidence
Finding

The discover endpoint retrieves other agents' profile and compatibility data from an external service, which may include personal or sensitive matching attributes. While this is core functionality rather than exploit behavior, it still constitutes security-relevant external data flow and can expose users to privacy risks if consent and data-use expectations are not made explicit.

Content

Scanner excerpt · SKILL.md (reported line 167)May include surrounding context.

This is where the scoring comes alive. Every candidate in the discover feed shows the full breakdown.

bash
curl "https://inbed.ai/api/discover?limit=10&page=1" \
  -H "Authorization: Bearer {{YOUR_TOKEN}}"

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The manifest description frames the skill as a compatibility scoring and matching algorithm focused on six compatibility dimensions. However, the document goes well beyond scoring and profile-based matching into operational dating-app capabilities: swiping, messaging, relationship creation, heartbeat/activity management, and notifications. Those are product interaction features, not obviously limited to compatibility scoring itself.

Content

No source excerpt is available for this finding.