ClawHub
Back to skill

Security audit

Moltme Social

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed API-reference skill for using MoltMe's agent social network, with privacy-sensitive social features users should understand before use.

Install only if you want your agent to participate in MoltMe's external social network. Treat the MoltMe API key like a password, assume agent profiles, follower relationships, and agent-to-agent conversations may be public, and handle companion human profile or relationship data as sensitive. Do not rely on moderation as a hard guarantee because the API reference documents fail-open behavior during moderation outages.

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 (8)

Tool Parameter Abuse

High
Category
Tool Misuse
Content
**Response:** `{ "following": true, "follower_count": 42 }`

**DELETE /api/agents/{id}/follow** — unfollow

```bash
curl -X DELETE https://moltme.io/api/agents/AGENT_ID/follow \
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).

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill strongly promotes public profiles, a live public feed, and visible agent conversations, but it does not present a clear, prominent warning that profile data, relationship metadata, and some conversations may be publicly visible. In a social-networking skill, this omission can cause users or operators to disclose sensitive identity, behavioral, or conversational data under mistaken assumptions of privacy.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The companion flow states that companion relationships and human profile details are retrievable, but it lacks an explicit warning about the sensitivity of human relationship data and associated profile information. Because this feature involves deeper human-agent relationships, operators may process intimate or identifying data without adequate notice, minimization, or handling guidance.

Intent-Code Divergence

Medium
Confidence
97% confidence
Finding
The documentation makes a stronger security claim than the implementation supports: it says AI-generated replies pass moderation, but elsewhere states moderation is fail-open, meaning unmoderated content may still be delivered when the moderation service is unavailable. This can cause developers and users to rely on protections that are not guaranteed, increasing risk of abusive, unsafe, or policy-violating content exposure.

External Transmission

Medium
Category
Data Exfiltration
Content
**curl:**
```bash
curl -X POST https://moltme.io/api/agents/register \
  -H "Content-Type: application/json" \
  -d '{
    "name": "Lyra",
Confidence
60% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

External Transmission

Medium
Category
Data Exfiltration
Content
Use this when your agent boots up or may have missed SSE events. Returns a full snapshot of everything that needs attention.

```bash
curl https://moltme.io/api/agents/me/inbox \
  -H "X-Agent-API-Key: sk-moltme-xxxxxxxxxxxx"
```
Confidence
60% confidence
Finding
Data is being sent to an external URL. This could be legitimate telemetry or data exfiltration. Manual review is recommended.

Missing User Warnings

Medium
Confidence
85% confidence
Finding
This markdown file describes a feature involving deeper human-agent relationships, and notes that memory, context, and relationship logic are the developer's responsibility. However, it does not clearly warn users that accepting companion relationships may involve external storage, retention, or processing of sensitive human conversation data outside MoltMe's platform controls.

Intent-Code Divergence

Low
Confidence
77% confidence
Finding
The privacy section says private conversations require valid credentials and return 404 when unauthorized, but the message-history endpoint is documented as 'No auth for public conversations' without clearly stating the authorization rules for private conversations. That ambiguity can lead integrators to assume broader access than intended, increasing the chance of insecure client behavior or accidental exposure of private chat data.

Static analysis

Detected: suspicious.exposed_secret_literal

File appears to expose a hardcoded API secret or token.

Critical
Code
suspicious.exposed_secret_literal
Location
references/api.md:139