Back to skill

Security audit

Agent Workspace Manager

Security checks for vulnerabilities and agentic risk

Overview

This skill is a Markdown-only workspace organization guide with disclosed file-scanning and cleanup behavior, and no hidden execution, network access, or credential handling.

Before installing, be comfortable with a skill that can guide an agent to scan and reorganize workspace files. For bulk renames, archiving, pruning, or any change to the installed skill/reference documents, require a clear proposed diff or file list and explicit approval.

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
  • Rogue AgentSelf-Modification, Session Persistence
  • 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 (4)

Self-Modification

High
Category
Rogue Agent
Confidence
92% confidence
Finding

The skill instructs the agent to 'update this skill's structure reference,' which is a self-modification capability. Allowing a skill to alter its own guidance can weaken review controls, enable persistence of unsafe instructions, and create a path for instruction drift or hidden policy changes over time, especially in multi-agent sharing scenarios.

Content

Scanner excerpt · SKILL.md (reported line 134)May include surrounding context.

md
1. **Detect:** Identify documents that don't fit current categories
2. **Propose:** Suggest new directories or structure reorganization
3. **Implement:** Create new directories following naming conventions
4. **Document:** Update this skill's structure reference

**Example:** If `knowledge/` grows too large with mixed domain types, split into `knowledge/product/`, `knowledge/technical/`, `knowledge/business/`.

Natural-Language Policy Violations

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The naming rule uses Chinese-only example filenames as the prescribed 'DO' pattern and forbids alternative suffix-based naming, but the document does not state that this skill is limited to a Chinese-language environment or offer users a choice of locale. This creates a natural-language locale policy issue because it implicitly enforces a specific language convention on all users.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
84% confidence
Finding

This changelog includes natural-language examples such as _最终版, _修正版, and 神策宽表转JSON.sql, which embed a specific language/locale in the skill's guidance. Because the file does not clarify that multilingual naming is optional or context-specific, it may imply a locale-specific convention without user opt-in.

Content

No source excerpt is available for this finding.

Natural-Language Policy Violations

Low
Category
Not specified by scanner
Confidence
82% confidence
Finding

The SQL naming guidance uses Chinese filenames as the normative examples and marks alternative suffixed forms as disallowed, which can amount to an implicit language/locale requirement. The document does not state that this convention is limited to a Chinese-language environment or offer any user choice on naming language.

Content

No source excerpt is available for this finding.