ClawHub
Back to skill

Security audit

EVR Framework

Security checks for vulnerabilities and agentic risk

Overview

This skill has a coherent verification purpose, but it broadly pushes agents to perform real actions and persist evidence or behavioral rules without clear user consent.

Install only if you want agents to be more forceful about executing and verifying tasks. Before using it, ensure your agent still requires explicit user approval for destructive actions, external messages, scheduled jobs, credentialed operations, and any persistent memory or WORK_RULES.md changes.

A.I.G

Vulnerability Patterns
  • Agent Memory PoisoningWrites attacker-controlled rules into memory that affect later sessions
  • Skill Instruction HijackingAlters the agent's session goals or safety constraints when the skill loads
  • 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
Findings (1)

T02 · Agent Memory Poisoning

Warning
Location
SKILL.md:136
Finding

Persistent Modification of Agent Memory and Behavioral Rules

Content
View full analysis
Remediation
View remediation

SkillSpector

Vulnerability Patterns
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • 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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Vague Triggers

High
Category
Not specified by scanner
Confidence
97% confidence
Finding

The skill’s activation criteria are very broad, including common phrases like 'done', 'finished', 'verify', and general skepticism about completion. In an agentic environment, this can cause the skill to activate during ordinary conversations and push the agent toward real-world actions or verification steps that were not explicitly requested, increasing the chance of unintended side effects.

Content

No source excerpt is available for this finding.

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The skill explicitly encourages real execution of sensitive actions such as writing files, executing shell commands, creating cron jobs, and sending messages, but it does not require consent, sandboxing, scope limitation, or risk disclosure. Because the whole purpose of the skill is to force actual execution rather than simulated completion, misuse or accidental activation could directly change system state, expose data, or trigger unauthorized communications.

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 181)May include surrounding context.

md
| Pattern | Why It's Wrong | Fix |
|---------|----------------|-----|
| "I've created X" | No verification | Run command to check X exists |
| "Done ✅" | No evidence | Attach verification output |
| "Should be working" | Uncertainty | Verify it's working |
| "Successfully completed" | Self-assertion | Show external confirmation |