Back to skill

Security audit

file-classification-manager

Security checks for vulnerabilities and agentic risk

Overview

This file-organizing skill has a coherent purpose, but it can broadly move workspace files and create directories without confirmation or strong path safety checks.

Review carefully before installing. Only use this skill in a disposable or backed-up workspace, avoid running cleanup_workspace_root on a real project, and do not pass untrusted project names or file paths until validation, dry-run confirmation, collision-safe moves, and rollback behavior are added.

Vulnerability Patterns
  • Insecure Skill Coding PracticesFinds exploitable flaws such as hardcoded secrets or command injection
  • 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
Findings (3)

T09 · Insecure Skill Coding Practices

Error
Location
index.js:11
Finding

Path Traversal Through Unvalidated Project Context

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
index.js:84
Finding

Destination File Can Be Overwritten During Routing

Content
View full analysis
Remediation
View remediation

T09 · Insecure Skill Coding Practices

Warning
Location
index.js:119
Finding

Workspace Cleanup Performs Broad Destructive Moves Without Confirmation

Content
View full analysis
Remediation
View remediation
Vulnerability Patterns
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • 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
Findings (7)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The description emphasizes automatic organization, directory creation, and cleanup behavior but does not prominently warn that the skill can modify the workspace by moving user files. Users or other agents may invoke it assuming it is advisory, when it is actually capable of changing file locations and project structure.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

Advertising legacy cleanup and migration without an explicit warning understates the operational risk of relocating existing files. In practice, this can disrupt user workflows, break references, or misclassify important files if the heuristics are wrong.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The skill is framed to activate during very common file-related workflows such as generating files, processing files via other skills, and organizing workspace clutter, but it does not define clear trigger boundaries or require explicit user approval before acting. In a file-moving skill, overly broad activation increases the chance of unintended file reorganization and surprising side effects across unrelated projects.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The cleanup API is described as scanning the workspace root and routing misplaced files, which is a broad modifying operation on existing content. Without a warning and confirmation gate, this creates a realistic risk of unintended file movement, project corruption, or loss of discoverability of user files.

Content

No source excerpt is available for this finding.

Autonomous Decision Making

Medium
Category
Excessive Agency
Confidence
75% confidence
Finding

Skill enables autonomous high-impact decisions without human-in-the-loop verification. Critical operations (destructive commands, financial transactions, data deletion) should require explicit user confirmation.

Content

Scanner excerpt · SKILL.md (reported line 112)May include surrounding context.

text

## Safety Rules
- Never delete files without confirmation
- Always backup before moving critical files  
- Preserve original file timestamps and metadata
- Log all file operations to `memory/YYYY-MM-DD.md`

Ssd 3

Medium
Category
Not specified by scanner
Confidence
89% confidence
Finding

Logging all file operations to persistent memory can retain sensitive filenames, paths, project names, timestamps, and metadata beyond the immediate task. This is especially risky in a classification skill because filenames and directory structures often reveal confidential subjects, clients, or internal project details.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The cleanup routine iterates over all non-directory files in the workspace root and automatically moves them based on inferred content, without any confirmation, dry-run mode, explicit allowlist, or user-visible warning. In an agent/workspace context this can unexpectedly relocate user data, interfere with other tooling, and cause loss of availability or confusion about file locations even if the files are not deleted.

Content

No source excerpt is available for this finding.