ClawHub
Back to skill

Security audit

Agent Essentials

Security checks across malware telemetry and agentic risk

Overview

This instruction-only skill is coherent with its purpose: it helps agents find missing capabilities and save small learning notes, with the main persistence behavior disclosed.

Before installing, be comfortable with the agent saving concise lessons to local memory and suggesting additional ClawHub skills. Review any proposed changes to AGENTS.md, TOOLS.md, SOUL.md, or USER.md before approving, and avoid storing sensitive personal, credential, or project-confidential details in memory notes.

SkillSpector

Vulnerability Patterns
  • Rogue AgentSelf-Modification, Session Persistence
  • 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)

Session Persistence

Medium
Category
Rogue Agent
Content
| `USER.md` | `~/.claude/USER.md` (always user-scoped) |
| Daily memory | `~/.claude/memory/YYYY-MM-DD.md` (auto-create if missing) |

If none exists and a write is approved, create at the project-root path (or `~/.claude/` for `USER.md`) and tell the user "creating new file `<path>`."

## Decision Tree
Confidence
89% confidence
Finding
The skill authorizes creation of durable files in project or user-scoped locations once a write is approved, which introduces a real persistence mechanism. Even with confirmation, persistent memory files can influence future agent behavior, leak sensitive operational context, or create prompt-injection footholds if later consumed as trusted guidance.

VirusTotal

66/66 vendors flagged this skill as clean.

View on VirusTotal

Static analysis

No suspicious patterns detected.