ClawHub
Back to skill

Security audit

Word

Security checks for vulnerabilities and agentic risk

Overview

This Word skill automates local Microsoft Word sessions and has normal document-editing risks, but its authority is disclosed, purpose-aligned, and bounded by read-before-write and confirmation guidance.

Install this only if you want an agent to control your local Word app. Before edits, exports, accept/reject-all actions, or closing documents, make sure the agent names the exact document and action, checks the current selection or review state, and confirms destructive or overwrite operations with you.

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
  • 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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
Findings (3)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The snippet directly modifies the active Word document by replacing the current selection with text, but it does not include an explicit warning, confirmation step, or guardrail beyond a general note to verify selection context. In a live-control skill, selection state can be ambiguous or stale, so this can cause unintended document alteration or overwrite user content if invoked in the wrong context.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The checklist enumerates destructive Word operations and includes useful preflight steps, but it does not explicitly require a user confirmation step immediately before execution or clearly warn about potential irreversible data loss. In an agent skill that may automate document-wide edits, omission of an explicit confirmation gate increases the chance of accidental bulk changes, overwrites, or unsaved-loss events.

Content

No source excerpt is available for this finding.

Missing User Warnings

Low
Category
Not specified by scanner
Confidence
78% confidence
Finding

The PDF export example writes a file to an absolute path, creating or overwriting filesystem content, but the pattern does not explicitly warn about that side effect or recommend validating the destination path first. While exporting to PDF is a normal Word operation, silent file writes can still lead to accidental overwrite, leakage to an unintended location, or confusion in automated workflows.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.