ClawHub
Back to skill

Security audit

Inngest

Security checks for vulnerabilities and agentic risk

Overview

This is a straightforward Inngest API helper, but it can affect real background jobs if given valid Inngest credentials.

Install this only if you want an agent to help operate your Inngest account. Use scoped and rotatable credentials where possible, avoid exposing keys in chat or logs, and require explicit confirmation of the environment, payload, and RUN_ID before sending events, canceling runs, or replaying failed jobs.

SkillSpector

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

Missing User Warnings

Medium
Confidence
91% confidence
Finding
The skill exposes a destructive cancellation operation against live workflow runs without any warning, confirmation guidance, or scope restrictions. In an agent context, this increases the chance of accidental or over-broad job termination, which can disrupt production workflows and cause loss of in-flight processing.

Static analysis

No suspicious patterns detected.