Openclaw Smart Cache

• Outcome: Pass
• Latest audit: May 29, 2026 at 8:25 PM UTC
• Version:
  0.1.0

Overview

This is a documented local caching and request-learning skill; its persistent logs are expected for that purpose, but users should avoid caching sensitive content.

Before installing, verify the external repository because this package only contains documentation. Treat the cache and learner as persistent local storage: do not record passwords, API keys, private personal data, or confidential business queries, and use the clear commands when stored history is no longer needed.

SkillSpector

By NVIDIA

Vulnerability Patterns

• Prompt Injection: Instruction Override, Hidden Instructions, Exfiltration Commands
• Data Exfiltration: External Transmission, Env Variable Harvesting, File System Enumeration
• Privilege Escalation: Excessive Permissions, Sudo/Root Execution, Credential Access
• Supply Chain: Unpinned Dependencies, External Script Fetching, Obfuscated Code
• Excessive Agency: Unrestricted Tool Access, Autonomous Decision Making, Scope Creep

Findings (5)

Missing User Warnings

Medium

• Confidence: 87% confidence
• Finding: The documentation exposes destructive clear commands for cache and learner data without any warning, confirmation, or indication that persisted state will be deleted. In an agent skill context, this can lead users or automation to erase operational data unintentionally, causing data loss and reducing auditability or reproducibility.

Missing User Warnings

Medium

• Confidence: 92% confidence
• Finding: The skill documents on-disk storage of patterns and request logs but provides no privacy warning or data-handling guidance, even though those artifacts may contain user queries and metadata. This omission can cause operators to unknowingly retain sensitive user-provided content on disk, increasing privacy, compliance, and local disclosure risks.

Ssd 3

Medium

• Confidence: 94% confidence
• Finding: The skill explicitly promotes automatic learning and recommendation features based on recording repeated requests, which naturally entails persistent retention of user request content and behavior patterns. In agent environments, such logs can accumulate sensitive prompts, business data, or identifiers and later be exposed through local files, reports, or debugging workflows.

Ssd 3

Medium

• Confidence: 95% confidence
• Finding: The examples instruct users to record full natural-language queries and expose recommendation/reporting flows, encouraging retention of raw user-provided content rather than privacy-preserving summaries. That increases the chance that sensitive prompts or personal/business information will be stored and later disclosed through logs, reports, or filesystem access.

Ssd 3

Medium

• Confidence: 93% confidence
• Finding: The documented file structure includes persisted patterns and request logs, and the CLI supports report/export behavior, together indicating that recorded user content may be stored and surfaced later. This makes the privacy risk more concrete because the data is not merely transient in memory but organized for later retrieval and possible disclosure.

VirusTotal

66/66 vendors flagged this skill as clean.