ClawHub
Back to skill

Security audit

Hubspot Implementation Plan

Security checks for vulnerabilities and agentic risk

Overview

This skill is a planning helper for HubSpot audit results, with no bundled executable code, but users should review CRM-changing follow-up tasks before acting on them.

Installers should treat this as a planning skill, not an execution safeguard. Before following any recommended CRM cleanup task, confirm the selected audit report, export affected records, review a dry run where available, and explicitly approve any delete, suppress, merge, reassignment, git push, or pull request action.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (4)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The skill recommends destructive actions such as deleting contacts, suppressing contacts, merging companies, and reassigning records without requiring explicit warnings, backup steps, dry-run output, or confirmation checkpoints. In a CRM context, these actions can cause irreversible data loss, billing-impacting state changes, and business process disruption if an audit is inaccurate or the wrong report is selected.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
87% confidence
Finding

The instructions explicitly direct use of the "Claude Anthropic Chrome Extension" as one workflow-building option. This names a specific external AI assistant/tool without framing it as user-selected opt-in language or offering a neutral choice at the point of instruction, which can violate language/locale-style organizational policy constraints around imposed assistant selection.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill's stated purpose is to generate an implementation plan, but this section extends behavior into creating skills, pushing to a fork, and opening pull requests. That scope expansion crosses from local planning into code generation and external publication actions, which increases the chance of unintended repository modification or exfiltration of internal process details if followed automatically.

Content

No source excerpt is available for this finding.

Description-Behavior Mismatch

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

This section explicitly instructs the agent to create new skills, add files under skills/, commit, push, and open a PR to an upstream repository. That is a material expansion from analysis/planning into code authoring and external side effects, which can be abused or triggered without sufficiently separate authorization boundaries.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.