ClawHub
Back to skill

Security audit

Convex

Security checks for vulnerabilities and agentic risk

Overview

This Convex skill is a disclosed local-memory guidance skill with no executable code, credential access, network calls, or hidden high-impact behavior.

Install only if you want persistent local Convex project notes under ~/convex/. Choose a narrow activation preference if you do not want the skill to engage on incidental Convex mentions, and avoid saving secrets, tokens, personal data, or sensitive tenant details in the memory files.

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
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
  • 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
  • Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code
Findings (3)

Credential Access

High
Category
Privilege Escalation
Content
- rollout notes for deployment and incident lessons
- auth notes for permission and boundary edge cases

Do not store secrets, access tokens, or personal data unless the user explicitly requests it.

## Core Rules
Confidence
70% confidence
Finding
Code accesses credential files (SSH keys, AWS credentials, etc.). This could indicate credential theft attempts.

Vague Triggers

Medium
Confidence
96% confidence
Finding
The setup instructs activation to be confirmed for "every time Convex is mentioned," which is overly broad and can cause the skill to engage outside clear user intent. This increases the chance of unnecessary context collection and persistence to local memory during incidental mentions, creating avoidable privacy and scope-creep risk.

Missing User Warnings

Low
Confidence
89% confidence
Finding
The skill directs saving activation preferences in ~/convex/memory.md without requiring a user-facing notice or consent flow. Even though it says not to store secrets, silent local persistence can still capture sensitive workflow, architecture, or auth-related context that the user may not expect to be written.

Static analysis

No suspicious patterns detected.