ClawHub
Back to skill

Security audit

Engrm Sentinel

Security checks for vulnerabilities and agentic risk

Overview

This skill is a small Markdown-only coding review aid that asks the agent to use existing Engrm context for risk prevention and does not add commands, persistence, or hidden access.

Installers should treat this as advisory review support: it may cause the agent to raise extra risk-prevention comments on security-sensitive coding tasks, but it does not grant new tools or ask the agent to run commands. Users who want tighter behavior may ask for clearer activation criteria before relying on it broadly.

A.I.G

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

SkillSpector

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
89% confidence
Finding

The README describes broad capabilities such as surfacing likely mistakes and applying lessons during risky coding work, but it does not define concrete activation criteria, boundaries, or decision rules for when the skill should be invoked. In an agent setting, this ambiguity can cause overbroad or mistimed activation, leading the skill to influence unrelated tasks, create noisy or misleading security guidance, or be used as a blanket authority without sufficient context.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
88% confidence
Finding

The phrase "when code is entering a risky area" is subjective and does not clearly define activation boundaries. Although later examples narrow it somewhat, the top-level trigger remains broad enough that different agents could invoke the skill inconsistently.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.