ClawHub

Core Skill

Security checks across malware telemetry and agentic risk

Overview

This skill is a disclosed agent identity and connection layer that requires explicit user approval for publishing, introductions, identity sharing, and other external actions.

Install only if you want an agent identity, signed receipts, and an external connection layer. Treat generated keys and any tokens as sensitive, review the npm packages and remote MCP endpoint, and approve each publish, search, introduction, delegation, or commerce-related action deliberately.

SkillSpector

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

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
## What this is not

This does not expand the agent's authority, autonomy, tool access, or
ability to act without approval. Authority is delegated by the principal
and only narrows at each transfer; revoking the principal's grant kills the
agent's downstream authority in one call. The agent does not publish a
network presence, request an introduction, or share the principal's
Confidence
75% confidence
Finding
without approval

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
get_digest

# 2. With explicit approval, describe what the principal needs/offers
publish_intent_card        # never without approval; never auto-published

# 3. Surface relevant matches for the principal to review
search_matches
Confidence
75% confidence
Finding
without approval

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
`respond_to_intro` without the principal's explicit approval for that
  specific action.
- Never share the principal's identity, contact details, or intent with
  another party without approval. Introductions exchange details only after
  both sides opt in.
- If nothing relevant is found, stay silent. Do not manufacture activity or
  nudge repeatedly.
Confidence
75% confidence
Finding
without approval

VirusTotal

VirusTotal findings are pending for this skill version.

View on VirusTotal