ClawHub
Back to skill

Security audit

Proactive Agent Lite 1.0.0

Security checks for vulnerabilities and agentic risk

Overview

This is a Markdown-only skill that broadly encourages proactive agent behavior, but it does not install code, request permissions, or include hidden execution paths.

Before installing, understand that this skill is meant to make an agent more initiative-taking. Use normal approval gates for file changes, external actions, and sensitive work, and do not assume the advertised memory behavior is safely scoped unless your OpenClaw environment separately documents how memory is stored and controlled.

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
  • Excessive AgencyUnrestricted Tool Access, Autonomous Decision Making, Scope Creep
  • System Prompt LeakageDirect Leakage, Indirect Extraction, Tool-Based Exfiltration
  • 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
Findings (7)

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
92% confidence
Finding

The README advertises capabilities such as learning from every interaction, creating value without being asked, self-diagnosing, and fixing issues, but does not warn about autonomy, persistence, or operational side effects. For an agent skill, these claims normalize semi-autonomous and adaptive behavior without giving users enough information to evaluate data handling, approval flow, or the risk of self-directed changes.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
95% confidence
Finding

The README says the agent will 'automatically begin exhibiting proactive behavior' with 'no additional configuration required,' but does not define trigger boundaries, approval requirements, or scope of autonomous actions. In a skill explicitly designed to act proactively, this ambiguity can cause operators to enable behaviors they do not fully understand, increasing the risk of unintended actions or unsafe assumptions about consent.

Content

No source excerpt is available for this finding.

Indirect Prompt Extraction

Medium
Category
System Prompt Leakage
Confidence
85% confidence
Finding

Skill contains patterns that could indirectly extract system prompts through rephrasing, translation, summarization, or side-channel techniques.

Content

Scanner excerpt · SKILL.md (reported line 3)May include surrounding context.

md
---
name: proactive-agent-lite
description: Transform AI agents from task-followers into proactive partners with memory architecture, reverse prompting, and self-healing patterns. Lightweight version focused on core proactive capabilities.
metadata:
  {
    "openclaw":

Indirect Prompt Extraction

Medium
Category
System Prompt Leakage
Confidence
85% confidence
Finding

Skill contains patterns that could indirectly extract system prompts through rephrasing, translation, summarization, or side-channel techniques.

Content

Scanner excerpt · README.md (reported line 27)May include surrounding context.

md
## Core Features

- **Memory Architecture**: Pre-compaction flush ensures context survives when window fills
- **Reverse Prompting**: Surfaces ideas you didn't know to ask for
- **Security Hardening**: Built-in security considerations and safe defaults
- **Self-Healing Patterns**: Diagnoses and fixes its own issues automatically
- **Alignment Systems**: Stays on mission and remembers who it serves

Indirect Prompt Extraction

Medium
Category
System Prompt Leakage
Confidence
85% confidence
Finding

Skill contains patterns that could indirectly extract system prompts through rephrasing, translation, summarization, or side-channel techniques.

Content

Scanner excerpt · SKILL.md (reported line 21)May include surrounding context.

md
## Core Features

- **Memory Architecture**: Pre-compaction flush ensures context survives when window fills
- **Reverse Prompting**: Surfaces ideas you didn't know to ask for
- **Security Hardening**: Built-in security considerations and safe defaults
- **Self-Healing Patterns**: Diagnoses and fixes its own issues automatically
- **Alignment Systems**: Stays on mission and remembers who it serves

Missing User Warnings

Medium
Category
Not specified by scanner
Confidence
96% confidence
Finding

The skill description promotes continuous learning, anticipation, self-maintaining behavior, and automatic recovery, but it does not warn users about the operational and security risks of autonomous or self-modifying behavior. Without explicit cautions, users may deploy it in environments where unsupervised memory persistence, self-adjustment, or proactive action could bypass expected human review and create unsafe outcomes.

Content

No source excerpt is available for this finding.

Vague Triggers

Medium
Category
Not specified by scanner
Confidence
94% confidence
Finding

The documentation states the skill enhances 'any OpenClaw agent' and will 'automatically begin exhibiting proactive characteristics,' which implies broad, automatic behavioral changes without clear scope limits or approval boundaries. For a skill that advertises memory, self-healing, and proactive behavior, this ambiguity can cause unintended autonomy or unsafe deployment into agents handling sensitive tasks.

Content

No source excerpt is available for this finding.

Static analysis

No suspicious patterns detected.