Security audit

Prepublish Privacy Scrub

Security checks for vulnerabilities and agentic risk

Overview

This skill is meant to remove sensitive data before publishing, but its scrubber can recursively overwrite many files in place and falsely claims backups are made.

Review this skill before installing or using it. Do not run Invoke-PrivacyScrub on a real project unless you have your own backup and have narrowed the target path, because the artifact's backup claim is not implemented and the scrubber may alter unintended files while still missing some sensitive data categories.

A.I.G

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 (2)

T09 · Insecure Skill Coding Practices

Warning

Location: SKILL.md:72
Finding: Recursive In-Place Scrubbing Can Corrupt Files Without Recovery
Content: View full analysis
Remediation: View remediation

T09 · Insecure Skill Coding Practices

Warning

Location: SKILL.md:27
Finding: Privacy Scrubber Does Not Remove Several Sensitive Data Classes It Claims to Address
Content: View full analysis
Remediation: View remediation

SkillSpector

Vulnerability Patterns

Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
MCP Tool PoisoningHidden Instructions, Unicode Deception, Parameter Description Injection
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 (3)

Missing User Warnings

Medium

Category: Not specified by scanner
Confidence: 92% confidence
Finding: The scrub function recursively enumerates files and overwrites them in place, but the markdown does not clearly warn users about the scope and destructive nature of those modifications. In a pre-publish workflow, this can lead to unintended bulk edits across a directory tree, potentially altering non-target files or damaging content without review.
Content: No source excerpt is available for this finding.

Intent-Code Divergence

Medium

Category: Not specified by scanner
Confidence: 97% confidence
Finding: The documentation claims original files are backed up before scrubbing, but the implementation overwrites files in place with Out-File and contains no backup logic. This creates a real integrity and recovery risk: users may permanently lose original content, including incorrectly redacted or corrupted files, because the behavior is safer on paper than in reality.
Content: No source excerpt is available for this finding.

Vague Triggers

Medium

Category: Not specified by scanner
Confidence: 88% confidence
Finding: The trigger section says to use the skill 'Before any skill publish,' 'Before git push of skills,' and 'After copying skill from working directory' without defining exclusions, exact scope, or negative examples. These conditions are broad enough to cause unintended invocation in routine workflows, especially 'after copying skill from working directory,' which is ambiguous.
Content: No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.