Back to skill

Security audit

Skill Debloater

Security checks for vulnerabilities and agentic risk

Overview

This skill is a coherent skill-maintenance helper that uses local read-only analysis scripts and requires user confirmation before deleting anything.

Install only if you want an agent to help reorganize skill files. Review any proposed automatic reorganizations, and only approve deletions after confirming the backup exists and the listed files are truly stale.

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
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (2)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
93% confidence
Finding

The description says to trigger when the user asks to "clean up old versions" or when a skill is "too long/verbose or flaky," which are broad conditions that can overlap with ordinary maintenance requests. It does not clearly bound what kinds of cleanup are in scope or give exclusion examples, increasing the chance of unintended invocation.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
91% confidence
Finding

The trigger text includes broad states like "too long," "burns too many tokens," and "behaves inconsistently," combined with generic requests like "clean up." These criteria lack precise thresholds or exclusion conditions, so agents may invoke the skill for many normal editing situations.

Content

No source excerpt is available for this finding.