ClawHub
Back to skill

Security audit

auto-reply-gated

Security checks for vulnerabilities and agentic risk

Overview

This skill drafts Instagram DM replies and keeps sending gated behind explicit human approval, with no evidence of automatic sending, credential access, or network exfiltration.

Before installing, understand that the skill stores draft recipient and message text locally and requires a human approval step before export. Use it only with message contents you are comfortable storing in local draft JSON files, and avoid passing unrelated files to the approve or export commands.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • 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 (1)

Autonomous Decision Making

Medium
Category
Excessive Agency
Content
- No browser cookie or session extraction.
- No CDP, WebSocket debugger, or private Instagram API access.
- No background watcher, polling daemon, or unattended cron reply.
- No approval bypass flags.
- No secret, token, or credential logging.

If a user asks to send, open Instagram's official UI, present the final draft again, and require the user to perform or explicitly approve the final send action.
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.

Static analysis

No suspicious patterns detected.