Back to skill

Security audit

excel-keep-format-edit

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed local Excel workflow for editing copied workbooks while preserving formatting, with no hidden code, persistence, or exfiltration found.

Install only if you need a Chinese-language Windows/Office workflow for local Excel COM edits. Before using it on financial statements, work on a copy, confirm the exact workbook, sheet, cells, and accounting assumptions, and avoid relying on the promotional external links for operation.

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
Vulnerability Patterns
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (4)

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
84% confidence
Finding

The skill is written entirely in Chinese and presents its name, triggers, and operating instructions only in that language, with no indication that users may choose another language or locale. Under the stated policy, forcing a specific language without user opt-in is a natural-language policy concern.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The trigger phrases are broad enough to match ordinary spreadsheet-editing requests, which can cause the agent to invoke this Windows/Excel-COM-specific workflow in contexts where it is unnecessary or inappropriate. In practice, that can lead to unsafe file handling, incorrect tool selection, or unintended modification of user spreadsheets when a simpler or safer method should have been used.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

The activation conditions describe broad business scenarios but do not specify sufficient guardrails, such as validating workbook type, confirming the target sheet, requiring backup/copy-only operation, or requiring explicit user authorization for financial statement edits. Because this skill edits financial reports and discusses manual adjustments, ambiguous activation increases the chance of accidental misuse, data integrity errors, or unauthorized changes to sensitive business records.

Content

No source excerpt is available for this finding.

Context-Inappropriate Capability

Low
Category
Not specified by scanner
Confidence
93% confidence
Finding

The manifest describes a local Windows/Office skill for preserving Excel formatting while editing financial report values. The author section adds WeChat contact info and multiple external web/image URLs, which are not justified by the operational purpose of editing spreadsheets and introduce unrelated external-facing capability/context.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.