growth-tracker

Security checks across malware telemetry and agentic risk

Overview

This skill is a local progress-tracking/reporting helper whose persistent metrics are disclosed and aligned with its purpose, though users should understand it keeps a local history.

Install only if you want OpenClaw to maintain a local history of your progress, task, learning, achievement, and goal metrics. Review ~/.openclaw/data/quantified_self.json periodically, use export/reset commands deliberately, and set auto_track to false or reduce retention_days if ongoing local profiling is not desired.

SkillSpector

By NVIDIA

Vulnerability Patterns

Data ExfiltrationExternal Transmission, Env Variable Harvesting, File System Enumeration
Trigger AbuseOverly Broad Trigger, Shadow Command Trigger, Keyword Baiting Trigger
Prompt InjectionInstruction Override, Hidden Instructions, Exfiltration Commands
Privilege EscalationExcessive Permissions, Sudo/Root Execution, Credential Access
Supply ChainUnpinned Dependencies, External Script Fetching, Obfuscated Code

Findings (8)

Vague Triggers

Medium

Confidence: 91% confidence
Finding: The basic-query trigger phrases are extremely broad and map to normal conversational requests such as asking how things are going or requesting a report. In a skill-routing system, this can cause unintended activation, leading the skill to collect, summarize, or expose tracked data when the user did not explicitly mean to invoke quantified-self behavior.

Vague Triggers

Medium

Confidence: 90% confidence
Finding: The advanced-query examples like '详细报告' and '成长趋势' are even more generic than the basic triggers and are likely to appear in ordinary dialogue. This increases the chance of accidental skill invocation and unintended access to historical analytics or retained activity data.

Missing User Warnings

Medium

Confidence: 88% confidence
Finding: The README documents auto-tracking and persistent local storage, but does not prominently disclose that user activity data may be collected and retained by default. This creates a privacy and consent issue: users may unknowingly generate a local behavioral log that can later be exported, reviewed, or exposed through accidental invocation.

Vague Triggers

Medium

Confidence: 89% confidence
Finding: The trigger conditions are very broad and can cause the skill to activate on ordinary conversations about progress, learning, or performance without clear user intent to enable tracking. In the context of a skill that persistently records interaction-derived data, over-triggering increases the chance of silent collection and retention of user-related activity data.

Missing User Warnings

Medium

Confidence: 95% confidence
Finding: The skill defines persistent local storage for detailed session metrics and later supports export/reset flows, but it does not require clear notice and consent before collection begins. Even if stored locally, this creates privacy risk because sensitive behavioral history may be accumulated without informed user awareness.

Ssd 3

Medium

Confidence: 86% confidence
Finding: The skill's core purpose is to quantify and retain broad categories of growth, tasks, learning, achievements, and efficiency from interactions, which implies collection of comprehensive interaction-derived data. Without strong minimization and consent boundaries, this can become a privacy-invasive profile of user and agent activity over time.

Ssd 3

High

Confidence: 98% confidence
Finding: These instructions explicitly direct continuous recording during conversations and saving full state after each round, creating ongoing session persistence of interaction-derived data. This is dangerous because it normalizes silent longitudinal tracking, increasing exposure of sensitive activity patterns and making later over-disclosure or misuse more likely.

Ssd 3

High

Confidence: 97% confidence
Finding: Providing complete stored data on request can expose the entire accumulated interaction profile in one step, including derived behavioral and performance history. If triggered accidentally, by ambiguous phrasing, or in a shared environment, this increases the risk of unintended disclosure of sensitive local records.

VirusTotal

64/64 vendors flagged this skill as clean.

View on VirusTotal