Back to skill

Security audit

Devin 192ac1

Security checks for vulnerabilities and agentic risk

Overview

This markdown-only skill is noisy and may trigger on overly broad phrases, but it does not request code execution, data access, persistence, or privileged authority.

Before installing, expect possible irrelevant activation because the skill mixes Devin AI platform notes with unrelated drawing and basketball video references. Tightening or removing generic triggers would improve routing, but the artifact does not show malicious or high-impact behavior.

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 (4)

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The trigger conditions are broad enough to activate on common mentions of 'Devin' or a generic product-description phrase, which can cause this skill to run when users are not actually requesting it. In this file, the risk is increased by the merged and noisy content, because accidental invocation could route unrelated conversations into irrelevant or confusing skill behavior.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

The trigger phrase '绘画教程' is extremely generic and overlaps with ordinary user requests about drawing tutorials, making unintended invocation likely. The surrounding skill content is especially mismatched—mixing Devin, drawing, sports, and merged sources—so accidental activation is more dangerous because it can hijack user intent and deliver irrelevant guidance.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

This repeated trigger section preserves the same generic phrase '绘画教程', so the collision risk remains and is compounded by duplication. Repeated broad triggers enlarge the attack surface for misrouting because multiple sections may independently match benign user requests that have nothing to do with this skill.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
97% confidence
Finding

This additional repeated trigger again activates on the common phrase '绘画教程', continuing the same unintended-invocation problem. Given the file's inconsistent merged content, accidental activation can mislead users, interfere with other skills, and reduce trust in the agent's routing behavior.

Content

No source excerpt is available for this finding.