ClawHub
Back to skill

Security audit

Linkedin Pipedream

Security checks across malware telemetry and agentic risk

Overview

This LinkedIn automation skill is purpose-aligned, but it needs review because it can use stored Pipedream credentials and hardcoded account details to post or delete public LinkedIn content.

Install only if you control the relevant LinkedIn, Pipedream, and organization accounts. Replace all hardcoded Telegram, member, organization, and auth provision IDs with verified values, avoid copying or committing anything from ~/.config/pdauth/config.json, and require explicit confirmation before posting, commenting, liking, or deleting content.

SkillSpector

By NVIDIA
Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (4)

Context-Inappropriate Capability

Medium
Confidence
94% confidence
Finding
The skill documentation instructs users to bypass the declared MCP interface and run custom Node.js code that pulls credentials from a local config file and uses hardcoded OAuth/account identifiers. That expands the skill's effective capability from normal LinkedIn actions into local secret handling and arbitrary code execution patterns, increasing the chance of credential exposure, misuse of Pipedream access, or unsafe copy/paste execution.

Context-Inappropriate Capability

Medium
Confidence
98% confidence
Finding
The documentation exposes real-looking personal and account-specific identifiers, including a member URN, external user ID, auth provision ID, and organization admin relationship. Even if not sufficient alone for compromise, these identifiers leak operational metadata, enable targeting, and can assist account correlation or social-engineering attacks against the linked user and organization.

Missing User Warnings

Medium
Confidence
87% confidence
Finding
The skill includes a delete-post capability with example commands but gives no confirmation, preview, or warning guidance despite being an irreversible destructive action in many workflows. In an automation context, this raises the risk of accidental content loss or misuse through prompt mistakes, especially because the surrounding instructions normalize direct execution.

Missing User Warnings

Medium
Confidence
92% confidence
Finding
The file lists sensitive local paths, including a credentials file, without warning that these locations may contain secrets and should not be read, copied, logged, or shared. In the context of an agent skill, exposing such paths can encourage downstream tools or users to access local secret material that is unrelated to ordinary LinkedIn posting tasks.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.