ClawHub
Back to skill

Security audit

talk-normal

Security checks for vulnerabilities and agentic risk

Overview

This skill does what it says: it installs a prompt into AGENTS.md to make agent replies more concise, with no network calls or credential use found.

Before installing, inspect the resolved AGENTS.md target and confirm you want an always-on writing-style prompt there. If you change your mind, run bash install.sh --uninstall from the skill directory to remove the marked block.

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
93% confidence
Finding
The installer writes to or creates AGENTS.md automatically, and will also remove and replace an existing managed block, without any confirmation or explicit path selection by the user. Because AGENTS.md changes agent behavior persistently, this can silently alter future LLM behavior in the current directory, home directory, or OPENCLAW_WORKSPACE, which is a meaningful security and trust boundary even if the script's stated purpose is legitimate.

Static analysis

No suspicious patterns detected.