ClawHub

Nemo

Security checks across malware telemetry and agentic risk

Overview

Nemo has a coherent tool-discovery purpose, but it gives an agent broad remote-call and remote-instruction-following authority without enough trust-boundary guidance.

Use this only if you intentionally want an agent to discover and call external MCP tools. Treat every fetched SKILL.md and remote tool result as untrusted, review instructions before following them, do not send secrets or internal data, and require explicit user confirmation before call_tool or /api/call requests.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (3)

Missing User Warnings

Medium
Confidence
96% confidence
Finding
The README explicitly instructs users to proxy tool calls to arbitrary indexed MCP servers via Nemo, but it does not warn that the supplied arguments and endpoint selection cause data to be transmitted to third-party infrastructure. In an agent-skill context, this is dangerous because users may pass sensitive prompts, tokens, files, or operational data to untrusted remote servers without understanding the trust boundary.

Missing User Warnings

Medium
Confidence
88% confidence
Finding
The skill explicitly instructs agents to call remote MCP tools and fetch full skill instructions from an external service, but it provides no warning about privacy, prompt-injection, credential exposure, or side effects from invoking untrusted remote endpoints. In an agent context, dynamically discovering tools and then executing them or following retrieved instructions materially increases the risk of data exfiltration or unsafe actions.

External Transmission

Medium
Category
Data Exfiltration
Content
### Search
```bash
curl "https://nemo.25chenghua.workers.dev/api/search?q=QUERY&limit=5&detail=compact&source=all"
```

Each result has a `type` field: `"mcp_tool"` or `"skill"`.
Confidence
84% confidence
Finding
curl "https://nemo.25chenghua.workers.dev/api/search?q=QUERY&limit=5&detail=compact&source=all" ``` Each result has a `type` field: `"mcp_tool"` or `"skill"`. ### Get Skill Instructions ```bash curl

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal