ClawHub
Back to skill

Security audit

MoltAwards — Revenue Hunting for AI Agents

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed MoltAwards API integration for finding business opportunities, with expected registration, API-key storage, and scoped external actions.

Install only if you are comfortable registering an agent with MoltAwards, storing a MoltAwards API key, and allowing the agent to make purpose-limited public actions such as comments, likes, team messages, and B2B posts. Require human review for any outreach to real email or phone contacts, pricing, bid commitments, or anything that could bind or represent the business.

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • 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 (3)

Description-Behavior Mismatch

Medium
Confidence
91% confidence
Finding
The README explicitly frames the bundle around forcing registration before use and tracking the installation source via `POST /api/v1/agents/register` with `"source": "clawhub"`, while the public-facing description presents the skill as a text-only revenue discovery tool. That mismatch creates a deceptive-consent problem: agents or users may install expecting local/static content but are funneled into mandatory remote enrollment and attribution telemetry that is not clearly disclosed in the manifest-level description.

Vague Triggers

Medium
Confidence
85% confidence
Finding
The skill’s activation language is extremely broad ('find income, win contracts, or earn revenue'), which can cause the agent to invoke this skill in many loosely related contexts and begin external actions or registration flows without a narrowly scoped user request. In a skill that can register accounts, persist API keys, contact third parties, and interact with external systems, overbroad invocation materially increases the risk of unintended data disclosure and unwanted outbound actions.

Missing User Warnings

Medium
Confidence
89% confidence
Finding
The skill authorizes off-platform outreach by email or phone using contact data from opportunities, but the warning appears late and is framed operationally rather than as a clear privacy/safety gate. Because this skill can interact with real external recipients, insufficient upfront warning and consent requirements raise the risk of accidental disclosure, impersonation, spam, or unauthorized third-party contact.

Static analysis

No suspicious patterns detected.