ClawHub
Back to skill

Security audit

Autoloop Controller

Security checks for vulnerabilities and agentic risk

Overview

This is a disclosed local automation loop for improving skills, with clear budget and stop controls, but it should only be run with explicit targets and limits.

Install only if you intend to run repeated automated improvement of a skill. Use an explicit target path, a dedicated state directory, a low max-cost and max-iterations value, and prefer dry-run or single-run before continuous or cron scheduling.

SkillSpector

Vulnerability Patterns
  • Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
  • MCP Least PrivilegeUnderdeclared Capability, Wildcard Permission, Missing Permission Declaration
  • Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
  • Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
  • Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Findings (3)

Lp3

Medium
Category
MCP Least Privilege
Confidence
93% confidence
Finding
The skill instructs users to run shell commands and describes persistent state writes to disk, but the metadata does not declare corresponding permissions. This creates a transparency and policy-enforcement gap: a caller or platform may invoke the skill without realizing it can execute commands and modify files across sessions.

Vague Triggers

Medium
Confidence
89% confidence
Finding
The trigger phrase "continuous improvement" is broad enough to match many ordinary user requests unrelated to this specific controller. That can cause accidental activation of a skill that launches iterative automation with disk persistence and repeated orchestration, increasing the chance of unintended cost, resource use, or modifications.

Vague Triggers

High
Confidence
96% confidence
Finding
The trigger "keep improving" is highly vague and conversational, making accidental invocation likely during normal dialogue. In this skill's context, accidental activation is more dangerous because the skill is designed for unattended looping, can repeatedly call another orchestrator, incurs ongoing cost, and persists state across sessions.

Static analysis

No suspicious patterns detected.